From 5037956a3c571a0fdf64c56bafbb2e2d9655e411 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 19 Oct 2026 07:39:30 +0000 Subject: [PATCH] docs(bench): record `proof find` against no-brief on the 10x graph Haiku answered every question on the 10x graph twice with `proof find` and twice without it, one fresh session per question. On the 245 comparable questions: - `find` got 231.1 and 229.1 right per run, against 210.9 and 210.7. Over both runs, it alone was right 48 times and no-brief alone 12 times (sign test p < 0.0001). - It read 61% fewer tokens and made 9 tool calls instead of 21. - "Unknown" answers fell from 32 and 35 per run to 14 and 12. - The gain held on the questions the cards were not tuned on, and with the seven timed-out sessions set aside. - On the 24 questions that name no Effort, `find` got 46 of 48 and no-brief 31 of 48. Every pass mark in the plan is met. The README gives the tables and what the runs do not show. A measurement Finding records the result in `.flatbread-proof/`, derived from the 10x Finding. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01DVVnhYL1Bz7mCfEKDoyLMW --- ...-and-cut-reading-on-t--ydqm2ybyqjedzw9y.md | 19 + benchmarks/proof-recall/README.md | 57 + .../proof-recall/runs/x10-find-r1-find.json | 2261 +++++++++++++++++ .../runs/x10-find-r1-nobrief.json | 2256 ++++++++++++++++ .../proof-recall/runs/x10-find-r2-find.json | 2259 ++++++++++++++++ .../runs/x10-find-r2-nobrief.json | 2255 ++++++++++++++++ 6 files changed, 9107 insertions(+) create mode 100644 .flatbread-proof/findings/fnd-search-cards-raise-accuracy-and-cut-reading-on-t--ydqm2ybyqjedzw9y.md create mode 100644 benchmarks/proof-recall/runs/x10-find-r1-find.json create mode 100644 benchmarks/proof-recall/runs/x10-find-r1-nobrief.json create mode 100644 benchmarks/proof-recall/runs/x10-find-r2-find.json create mode 100644 benchmarks/proof-recall/runs/x10-find-r2-nobrief.json diff --git a/.flatbread-proof/findings/fnd-search-cards-raise-accuracy-and-cut-reading-on-t--ydqm2ybyqjedzw9y.md b/.flatbread-proof/findings/fnd-search-cards-raise-accuracy-and-cut-reading-on-t--ydqm2ybyqjedzw9y.md new file mode 100644 index 00000000..14afde79 --- /dev/null +++ b/.flatbread-proof/findings/fnd-search-cards-raise-accuracy-and-cut-reading-on-t--ydqm2ybyqjedzw9y.md @@ -0,0 +1,19 @@ +--- +id: fnd-search-cards-raise-accuracy-and-cut-reading-on-t--ydqm2ybyqjedzw9y +effort: eff-effort-graph-memory-and-agent-wedge--szeqvmgqjqnhd002 +title: Search cards raise accuracy and cut reading on the 10x graph +kind: measurement +created_at: '2026-10-19T07:39:20.029Z' +derives_from: + - fnd-a-graph-ten-times-larger-hides-answers-from-both--f9h4wcrg4jpzczzg +--- + +`proof find` searches every record by what it says and returns short cards: the current version of each hit, its best passages, any attached Blob passage, and one hop of links. On the benchmark graph ten times larger, Haiku answered the 245 comparable questions twice with `find` and twice without it (no-brief), one fresh session per question. + +- Accuracy rose by about 20 questions per run: 231.1 and 229.1 with `find`, against 210.9 and 210.7 without. Over both runs, `find` alone was right 48 times and no-brief alone 12 times (sign test p < 0.0001). +- Agents read 61 percent fewer tokens (13,019 against 33,839 on average) and made 9 tool calls instead of 21. +- "Unknown" answers fell from 32 and 35 per run to 14 and 12. +- No question category lost ground. The gain held on the half of the questions the cards were not tuned on (26 wins, 6 losses), and with the seven timed-out sessions set aside (45 wins, 9 losses). +- On 24 questions that name no Effort, `find` got 46 of 48 and no-brief 31 of 48. + +So the limit at scale was finding a record by what it says, as the 10x Finding suggested, and a search read lifts it. Open: the `brief` arm was not rerun, the graph is fictional and its filler easy to tell apart, and only Haiku was tested. diff --git a/benchmarks/proof-recall/README.md b/benchmarks/proof-recall/README.md index b693cd58..f13f4e00 100644 --- a/benchmarks/proof-recall/README.md +++ b/benchmarks/proof-recall/README.md @@ -767,3 +767,60 @@ lookups, though, so they are easier than the full set, and they test subsystem names cost the gate one point of evidence and no answers, so the 10x result is not only a sign that filler is easy to tell apart. The paid runs (Runs A to C in the plan) still have to show that agents gain from it. + +### `proof find` against no-brief on the 10x graph + +The plan's Run A and Run B, on Haiku, twice: `run.mjs --strategies find,nobrief --only default,ambiguous --graph `, 540 sessions per +run. The `find` arm's task text starts with `proof find`, then `get` for a +hit's full text, and `records` only to count. Both runs used the cards above +(`find_hash` `e5be6532…`). Run files: `runs/x10-find-r1-*.json` and +`runs/x10-find-r2-*.json`. + +On the 245 original questions (all but `world-077`, whose count includes +filler), pooled over both runs: + +| Measure | `find` | no-brief | +| ------------------------------------------ | ------ | -------- | +| Correct, run 1 | 231.1 | 210.9 | +| Correct, run 2 | 229.1 | 210.7 | +| Runs of a question only this arm got right | 48 | 12 | +| Mean read tokens | 13,019 | 33,839 | +| Median read tokens | 8,888 | 29,680 | +| Mean tool calls | 9 | 21 | + +The sign test over the pooled disagreements gives p < 0.0001. Counting each +question once, `find` did better on 37 and worse on 10 (p = 0.0001). It read +fewer tokens than no-brief in 447 of 490 sessions. + +Against the plan's pass marks, on its 232 comparable questions (the 245 +less the 13 retraction questions): + +| Mark | Result | Pass | +| ------------------------------------------ | -------------------------------------------------- | ---- | +| (a) Pooled sign test p < 0.05 | 46 wins, 9 losses, p = 4e-7 | yes | +| (b) Read tokens at most 0.65 of no-brief's | 13,260 against 34,313: 0.39 | yes | +| (c) "Unknown" falls by at least 10 per run | 32 to 14 and 35 to 12 | yes | +| No category loses more than 2 questions | No category loses any; `negative` 30 of 32 in both | yes | + +Seven sessions hit the 10-minute limit, six of them in the no-brief arm. +Setting those questions aside leaves the result in place: 45 wins, 9 +losses. The gain also holds on the questions the cards were not tuned on: +26 wins and 6 losses on the holdout half (p = 0.0005), against 23 and 6 on +the tune half. The largest gains are in `lookup` (+13.8 over two runs) and +`truncation` (+13), where the fact sits past the 600-character cut that +`records` applies; the smallest are in `temporal`, `negative`, `rationale`, +and `dead-evidence` (0 to +1). + +**The ambiguous-Effort set (Run B).** On the 24 questions that name no +Effort, pooled: `find` 46 of 48, no-brief 31 of 48. `find` did better on 10 +questions and worse on none (p = 0.002), and read 9,914 tokens to +no-brief's 42,017. + +What this does not show: + +- The `brief` arm was not run, so this says nothing new about compaction. +- Haiku only, and on a fictional graph whose filler is easier to tell apart + than real records. Run C, on the `--filler-overlap 0.3` graph, tests that. +- The ambiguous questions come from tiers 0 to 2. They show that search + finds the right Effort, not that agents handle a hard question without a + named Effort. diff --git a/benchmarks/proof-recall/runs/x10-find-r1-find.json b/benchmarks/proof-recall/runs/x10-find-r1-find.json new file mode 100644 index 00000000..1c293787 --- /dev/null +++ b/benchmarks/proof-recall/runs/x10-find-r1-find.json @@ -0,0 +1,2261 @@ +{ + "run": { + "name": "x10-find-r1-find", + "protocol": "one fresh session per question", + "models": ["claude-haiku-5-5"], + "find_hash": "e5be653299abe9b2329822d98ffe09e90ee7cfe96f8711c813a51dc6500b887f" + }, + "answers": { + "ambiguous-001": { + "answer": "384 MB", + "tokens": 11381, + "fixed_tokens": 1107, + "processed_tokens": 78766, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-002": { + "answer": "99.6 percent", + "tokens": 7592, + "fixed_tokens": 1105, + "processed_tokens": 58464, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-003": { + "answer": "92.4%", + "tokens": 13502, + "fixed_tokens": 1109, + "processed_tokens": 123167, + "tool_calls": 11, + "turns": 10 + }, + "ambiguous-004": { + "answer": "6 min 40 s", + "tokens": 8193, + "fixed_tokens": 1104, + "processed_tokens": 59983, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-005": { + "answer": "6 weeks", + "tokens": 7985, + "fixed_tokens": 1102, + "processed_tokens": 59578, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-006": { + "answer": "16-bit float per channel", + "tokens": 13753, + "fixed_tokens": 1105, + "processed_tokens": 82824, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-007": { + "answer": "32 steps", + "tokens": 11013, + "fixed_tokens": 1107, + "processed_tokens": 77308, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-008": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 11443, + "fixed_tokens": 1107, + "processed_tokens": 67119, + "tool_calls": 9, + "turns": 6 + }, + "ambiguous-009": { + "answer": "3.0 ms", + "tokens": 13308, + "fixed_tokens": 1102, + "processed_tokens": 81255, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-010": { + "answer": "160 MB/s", + "tokens": 9621, + "fixed_tokens": 1103, + "processed_tokens": 62644, + "tool_calls": 7, + "turns": 6 + }, + "ambiguous-011": { + "answer": "260 MB", + "tokens": 12970, + "fixed_tokens": 1104, + "processed_tokens": 108307, + "tool_calls": 8, + "turns": 9 + }, + "ambiguous-012": { + "answer": "11 m", + "tokens": 11027, + "fixed_tokens": 1106, + "processed_tokens": 77310, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-013": { + "answer": "61%", + "tokens": 7807, + "fixed_tokens": 1108, + "processed_tokens": 59303, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-014": { + "answer": "textures", + "tokens": 7868, + "fixed_tokens": 1106, + "processed_tokens": 59223, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-015": { + "answer": "unknown", + "tokens": 22870, + "fixed_tokens": 1106, + "processed_tokens": 224942, + "tool_calls": 13, + "turns": 12 + }, + "ambiguous-016": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 7930, + "fixed_tokens": 1102, + "processed_tokens": 58629, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 10215, + "fixed_tokens": 1110, + "processed_tokens": 64430, + "tool_calls": 7, + "turns": 6 + }, + "ambiguous-018": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 8072, + "fixed_tokens": 1109, + "processed_tokens": 59883, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-019": { + "answer": "7.4 ms", + "tokens": 8025, + "fixed_tokens": 1107, + "processed_tokens": 59485, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-020": { + "answer": "5912 pages", + "tokens": 9325, + "fixed_tokens": 1104, + "processed_tokens": 62475, + "tool_calls": 7, + "turns": 6 + }, + "ambiguous-021": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 9173, + "fixed_tokens": 1103, + "processed_tokens": 85555, + "tool_calls": 7, + "turns": 8 + }, + "ambiguous-022": { + "answer": "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "tokens": 17414, + "fixed_tokens": 1105, + "processed_tokens": 96786, + "tool_calls": 8, + "turns": 7 + }, + "ambiguous-023": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 11216, + "fixed_tokens": 1108, + "processed_tokens": 81130, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-024": { + "answer": "fnd-streaming-tide-heights-costs-46-kb-s-per-peer--knxqtyqh6xcz1tp9", + "tokens": 11106, + "fixed_tokens": 1102, + "processed_tokens": 83138, + "tool_calls": 6, + "turns": 7 + }, + "cross-001": { + "answer": "con-frame-budget--br0qszfg8fys908n", + "tokens": 7593, + "fixed_tokens": 1096, + "processed_tokens": 58733, + "tool_calls": 5, + "turns": 6 + }, + "cross-002": { + "answer": 2, + "tokens": 12440, + "fixed_tokens": 1098, + "processed_tokens": 67923, + "tool_calls": 5, + "turns": 6 + }, + "cross-003": { + "answer": 2, + "tokens": 10198, + "fixed_tokens": 1096, + "processed_tokens": 73711, + "tool_calls": 8, + "turns": 7 + }, + "cross-004": { + "answer": "dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e", + "tokens": 9721, + "fixed_tokens": 1102, + "processed_tokens": 87618, + "tool_calls": 7, + "turns": 8 + }, + "cross-005": { + "answer": "Rhea Valdés", + "tokens": 13690, + "fixed_tokens": 1096, + "processed_tokens": 126054, + "tool_calls": 9, + "turns": 10 + }, + "cross-006": { + "answer": "19%", + "tokens": 8661, + "fixed_tokens": 1095, + "processed_tokens": 61335, + "tool_calls": 5, + "turns": 6 + }, + "cross-007": { + "answer": "8 GB", + "tokens": 10664, + "fixed_tokens": 1097, + "processed_tokens": 117579, + "tool_calls": 9, + "turns": 10 + }, + "cross-008": { + "answer": "unknown", + "tokens": 24425, + "fixed_tokens": 1092, + "processed_tokens": 308727, + "tool_calls": 15, + "turns": 16 + }, + "cross-009": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 12006, + "fixed_tokens": 1097, + "processed_tokens": 94040, + "tool_calls": 7, + "turns": 8 + }, + "cross-010": { + "answer": "dec-build-an-on-device-pipeline-cache-for-skiff--726r9mtk76qndeh7", + "tokens": 8144, + "fixed_tokens": 1095, + "processed_tokens": 59743, + "tool_calls": 5, + "turns": 6 + }, + "cross-011": { + "answer": false, + "tokens": 9576, + "fixed_tokens": 1094, + "processed_tokens": 87479, + "tool_calls": 7, + "turns": 8 + }, + "cross-012": { + "answer": "61 MB/s", + "tokens": 7770, + "fixed_tokens": 1097, + "processed_tokens": 58719, + "tool_calls": 5, + "turns": 6 + }, + "cross-013": { + "answer": "Skiff runs Tern at the Performance preset, saving 1.8 ms over Quality (2.1 ms vs 3.9 ms), per dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3. The Meridian 120 Hz mode's lowest preset is Ultra Performance, which Skiff does not use.", + "tokens": 17166, + "fixed_tokens": 1102, + "processed_tokens": 108978, + "tool_calls": 9, + "turns": 8 + }, + "cross-014": { + "answer": 3, + "tokens": 4800, + "fixed_tokens": 1104, + "processed_tokens": 39298, + "tool_calls": 4, + "turns": 5 + }, + "cross-015": { + "answer": "Two Efforts are no longer active. \"Co-op rollback netcode prototype\" was abandoned: rollback re-runs tides every frame, costing 4.8 ms on Skiff, so co-op will ship without it (dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b). \"Shader compile cache and build farm\" was completed once every farm build read from the cache (fnd-retrospective-on-the-shader-compile-cache--9rpv2a60ach04cx6, which reports the cache running on every farm build with a 94% hit rate and clean rebuilds down from 47 minutes to 6 min 40 s).", + "tokens": 22073, + "fixed_tokens": 1099, + "processed_tokens": 196924, + "tool_calls": 11, + "turns": 12 + }, + "cross-016": { + "answer": "16.7 ms", + "tokens": 9367, + "fixed_tokens": 1105, + "processed_tokens": 62861, + "tool_calls": 7, + "turns": 6 + }, + "cross-017": { + "answer": "false: dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e", + "tokens": 13672, + "fixed_tokens": 1096, + "processed_tokens": 82001, + "tool_calls": 6, + "turns": 7 + }, + "netcode-001": { + "answer": "abandoned", + "tokens": 8968, + "fixed_tokens": 1106, + "processed_tokens": 61730, + "tool_calls": 5, + "turns": 6 + }, + "netcode-002": { + "answer": "netcode/rollback-archive", + "tokens": 7945, + "fixed_tokens": 1098, + "processed_tokens": 59209, + "tool_calls": 5, + "turns": 6 + }, + "netcode-003": { + "answer": "38 KB", + "tokens": 8084, + "fixed_tokens": 1100, + "processed_tokens": 60007, + "tool_calls": 5, + "turns": 6 + }, + "netcode-004": { + "answer": "Rollback was abandoned because co-op needs bit-for-bit deterministic tides, and rollback must re-run the tide simulation every rolled-back frame, costing 4.8 ms on Skiff (a fifth of the 25 ms frame); the reason depends on the world Decision dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e (abandon decision: dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b).", + "tokens": 11608, + "fixed_tokens": 1102, + "processed_tokens": 78000, + "tool_calls": 6, + "turns": 7 + }, + "netcode-005": { + "answer": "unknown", + "tokens": 10996, + "fixed_tokens": 1103, + "processed_tokens": 89430, + "tool_calls": 7, + "turns": 8 + }, + "netcode-006": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 11822, + "fixed_tokens": 1098, + "processed_tokens": 118906, + "tool_calls": 9, + "turns": 10 + }, + "netcode-007": { + "answer": "dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b", + "tokens": 8123, + "fixed_tokens": 1099, + "processed_tokens": 59702, + "tool_calls": 5, + "turns": 6 + }, + "netcode-008": { + "answer": "dec-use-a-4-frame-rollback-window--pqe4hvbgmbp0dmh9", + "tokens": 13437, + "fixed_tokens": 1099, + "processed_tokens": 110206, + "tool_calls": 8, + "turns": 9 + }, + "netcode-009": { + "answer": "1.9 ms", + "tokens": 10984, + "fixed_tokens": 1098, + "processed_tokens": 117291, + "tool_calls": 9, + "turns": 10 + }, + "netcode-010": { + "answer": "dec-use-a-6-frame-rollback-window-with-1-frame-of-in--s53944zhkw103frk", + "tokens": 12872, + "fixed_tokens": 1104, + "processed_tokens": 83675, + "tool_calls": 8, + "turns": 7 + }, + "netcode-011": { + "answer": "con-co-op-peers-must-agree-on-tide-state-bit-for-bit--4v0sdbc3t4ywzr3g", + "tokens": 18216, + "fixed_tokens": 1097, + "processed_tokens": 241360, + "tool_calls": 15, + "turns": 16 + }, + "netcode-012": { + "answer": "dec-re-simulate-tides-at-quarter-resolution-during-r--7ah25p0r0xt9719f", + "tokens": 15314, + "fixed_tokens": 1104, + "processed_tokens": 161561, + "tool_calls": 11, + "turns": 12 + }, + "netcode-013": { + "answer": [ + "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "fnd-rollback-re-simulates-a-4-frame-window-in-1-1-ms--xa30a1cez2brphyf", + "rsk-re-simulation-cost-could-blow-the-skiff-frame-bu--sk2njsp170a3qwyf" + ], + "tokens": 26276, + "fixed_tokens": 1102, + "processed_tokens": 221064, + "tool_calls": 11, + "turns": 12 + }, + "netcode-014": { + "answer": [ + "iss-late-join-during-a-tide-surge-desyncs-the-joinin--mya93va0n7s8g7cj" + ], + "tokens": 44949, + "fixed_tokens": 1104, + "processed_tokens": 537467, + "tool_calls": 18, + "turns": 19 + }, + "netcode-015": { + "answer": 2, + "tokens": 8667, + "fixed_tokens": 1104, + "processed_tokens": 86478, + "tool_calls": 8, + "turns": 9 + }, + "ports-001": { + "answer": "25 ms", + "tokens": 9145, + "fixed_tokens": 1104, + "processed_tokens": 62198, + "tool_calls": 7, + "turns": 6 + }, + "ports-002": { + "answer": "384 MB", + "tokens": 8182, + "fixed_tokens": 1091, + "processed_tokens": 59900, + "tool_calls": 5, + "turns": 6 + }, + "ports-003": { + "answer": "61 MB/s", + "tokens": 8073, + "fixed_tokens": 1091, + "processed_tokens": 59623, + "tool_calls": 5, + "turns": 6 + }, + "ports-004": { + "answer": "accepted", + "tokens": 7366, + "fixed_tokens": 1098, + "processed_tokens": 57506, + "tool_calls": 5, + "turns": 6 + }, + "ports-005": { + "answer": "unknown", + "tokens": 20895, + "fixed_tokens": 1095, + "processed_tokens": 161182, + "tool_calls": 9, + "turns": 10 + }, + "ports-034": { + "answer": "6 weeks", + "tokens": 7946, + "fixed_tokens": 1095, + "processed_tokens": 59377, + "tool_calls": 5, + "turns": 6 + }, + "ports-006": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 15507, + "fixed_tokens": 1096, + "processed_tokens": 123295, + "tool_calls": 8, + "turns": 9 + }, + "ports-007": { + "answer": "SKF-UI-031", + "tokens": 7607, + "fixed_tokens": 1096, + "processed_tokens": 58440, + "tool_calls": 5, + "turns": 6 + }, + "ports-008": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 9196, + "fixed_tokens": 1091, + "processed_tokens": 85536, + "tool_calls": 7, + "turns": 8 + }, + "ports-009": { + "answer": "dec-load-skiff-fast-travel-without-a-loading-screen--v4qft5znysjv5f24", + "tokens": 8066, + "fixed_tokens": 1100, + "processed_tokens": 59493, + "tool_calls": 5, + "turns": 6 + }, + "ports-010": { + "answer": "con-frame-budget--wt5rfqh5z444g2cg", + "tokens": 15244, + "fixed_tokens": 1100, + "processed_tokens": 107402, + "tool_calls": 9, + "turns": 8 + }, + "ports-011": { + "answer": "99.6 percent", + "tokens": 7608, + "fixed_tokens": 1097, + "processed_tokens": 58353, + "tool_calls": 5, + "turns": 6 + }, + "ports-012": { + "answer": "They raised the Skiff dynamic resolution floor to 576p because saltglass caustics shimmered badly at the 540p floor, even after Tern (visible in 7 of 10 dune views, versus 2 of 10 at 576p). See dec-raise-the-skiff-dynamic-resolution-floor-to-576p--fep9q7m08pp9f0bb, citing fnd-saltglass-caustics-shimmer-at-the-540p-floor--0aa8sts89vhq1s6q.", + "tokens": 9345, + "fixed_tokens": 1096, + "processed_tokens": 86030, + "tool_calls": 7, + "turns": 8 + }, + "ports-013": { + "answer": "Ship Tern as Lantern's only upscaler", + "tokens": 9987, + "fixed_tokens": 1094, + "processed_tokens": 89175, + "tool_calls": 7, + "turns": 8 + }, + "ports-014": { + "answer": "dec-scale-skiff-resolution-dynamically-between-540p--jtytvgh2erd32x68", + "tokens": 10056, + "fixed_tokens": 1094, + "processed_tokens": 89510, + "tool_calls": 7, + "turns": 8 + }, + "ports-015": { + "answer": "textures", + "tokens": 9214, + "fixed_tokens": 1088, + "processed_tokens": 86002, + "tool_calls": 7, + "turns": 8 + }, + "ports-016": { + "answer": "dec-ship-prebuilt-pipeline-blobs-in-the-skiff-packag--3vjysjgy9qkvzebn", + "tokens": 11612, + "fixed_tokens": 1098, + "processed_tokens": 94410, + "tool_calls": 7, + "turns": 8 + }, + "ports-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 8320, + "fixed_tokens": 1096, + "processed_tokens": 60319, + "tool_calls": 5, + "turns": 6 + }, + "ports-018": { + "answer": "504p", + "tokens": 7980, + "fixed_tokens": 1099, + "processed_tokens": 59602, + "tool_calls": 5, + "turns": 6 + }, + "ports-019": { + "answer": "4 MB per minute", + "tokens": 9926, + "fixed_tokens": 1094, + "processed_tokens": 112561, + "tool_calls": 9, + "turns": 10 + }, + "ports-020": { + "answer": "fnd-skiff-sdk-1-2-takes-1-gb-back-from-games--a0ybdgjzw64hwfqa", + "tokens": 10744, + "fixed_tokens": 1098, + "processed_tokens": 92334, + "tool_calls": 9, + "turns": 8 + }, + "ports-021": { + "answer": "SKF-DSP-004", + "tokens": 9062, + "fixed_tokens": 1100, + "processed_tokens": 85590, + "tool_calls": 7, + "turns": 8 + }, + "ports-022": { + "answer": "cit-skiff-sdk-1-2-release-notes--7gwwn8xxffrx6npe", + "tokens": 11794, + "fixed_tokens": 1102, + "processed_tokens": 123162, + "tool_calls": 9, + "turns": 10 + }, + "ports-023": { + "answer": "fnd-tern-costs-3-9-ms-at-quality-and-2-1-ms-at-perfo--dkbskbqfgfwtryq0", + "tokens": 11168, + "fixed_tokens": 1099, + "processed_tokens": 129606, + "tool_calls": 10, + "turns": 11 + }, + "ports-024": { + "answer": "fnd-skiff-firmware-1-2-rejects-pipeline-blobs-built--pernvgb2he7m3z3y", + "tokens": 11643, + "fixed_tokens": 1091, + "processed_tokens": 123535, + "tool_calls": 9, + "turns": 10 + }, + "ports-025": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 9444, + "fixed_tokens": 1094, + "processed_tokens": 87646, + "tool_calls": 7, + "turns": 8 + }, + "ports-026": { + "answer": "912 MB", + "tokens": 12875, + "fixed_tokens": 1094, + "processed_tokens": 155412, + "tool_calls": 11, + "turns": 12 + }, + "ports-027": { + "answer": [ + "dec-stream-skiff-world-tiles-in-2-mb-reads--3jz9w9gvhhc7eppw", + "rsk-tide-streaming-outruns-skiff-storage--6q8s3b383z2tyeyb" + ], + "tokens": 45501, + "fixed_tokens": 1105, + "processed_tokens": 447871, + "tool_calls": 17, + "turns": 18 + }, + "ports-028": { + "answer": 2, + "tokens": 7715, + "fixed_tokens": 1100, + "processed_tokens": 62060, + "tool_calls": 6, + "turns": 7 + }, + "ports-029": { + "answer": null, + "tokens": 85863, + "fixed_tokens": 1098, + "processed_tokens": 3534567, + "tool_calls": 68, + "turns": 68 + }, + "ports-030": { + "answer": [ + "dec-cap-tide-storm-particles-at-6-000-on-skiff--resze52afjxyxys1", + "dec-drop-skiff-to-30-hz-while-a-tide-storm-is-on-scr--72vx1c8s4jrs6axc" + ], + "tokens": 7584, + "fixed_tokens": 1098, + "processed_tokens": 55751, + "tool_calls": 5, + "turns": 6 + }, + "ports-031": { + "answer": [ + "iss-meridian-cert-activity-cards-missing-for-two-bio--k214cmfra7pyvvzq", + "iss-skiff-cert-docked-30-hz-judder-on-some-tvs--9nq2xg9mzash474g" + ], + "tokens": 18375, + "fixed_tokens": 1096, + "processed_tokens": 157458, + "tool_calls": 12, + "turns": 11 + }, + "ports-032": { + "answer": "unknown", + "tokens": 7034, + "fixed_tokens": 1100, + "processed_tokens": 62923, + "tool_calls": 6, + "turns": 7 + }, + "ports-033": { + "answer": "unknown", + "tokens": 9823, + "fixed_tokens": 1092, + "processed_tokens": 60319, + "tool_calls": 5, + "turns": 6 + }, + "render-001": { + "answer": "7.4 ms", + "tokens": 9205, + "fixed_tokens": 1093, + "processed_tokens": 85954, + "tool_calls": 7, + "turns": 8 + }, + "render-002": { + "answer": "3.4x", + "tokens": 7567, + "fixed_tokens": 1092, + "processed_tokens": 58208, + "tool_calls": 5, + "turns": 6 + }, + "render-003": { + "answer": "Tomasz Brandt", + "tokens": 9462, + "fixed_tokens": 1097, + "processed_tokens": 87171, + "tool_calls": 7, + "turns": 8 + }, + "render-004": { + "answer": "rejected", + "tokens": 8017, + "fixed_tokens": 1090, + "processed_tokens": 59602, + "tool_calls": 5, + "turns": 6 + }, + "render-005": { + "answer": true, + "tokens": 7603, + "fixed_tokens": 1092, + "processed_tokens": 57917, + "tool_calls": 5, + "turns": 6 + }, + "render-006": { + "answer": "unknown", + "tokens": 8479, + "fixed_tokens": 1094, + "processed_tokens": 61435, + "tool_calls": 6, + "turns": 6 + }, + "render-007": { + "answer": "2026-07-24", + "tokens": 14554, + "fixed_tokens": 1093, + "processed_tokens": 85163, + "tool_calls": 6, + "turns": 7 + }, + "render-008": { + "answer": "2 failures", + "tokens": 7523, + "fixed_tokens": 1095, + "processed_tokens": 57958, + "tool_calls": 5, + "turns": 6 + }, + "render-009": { + "answer": "32 steps", + "tokens": 7869, + "fixed_tokens": 1099, + "processed_tokens": 58898, + "tool_calls": 5, + "turns": 6 + }, + "render-010": { + "answer": "unknown", + "tokens": 14412, + "fixed_tokens": 1094, + "processed_tokens": 105116, + "tool_calls": 7, + "turns": 8 + }, + "render-011": { + "answer": "fnd-sdf-gi-costs-2-4-ms-on-meridian--2w6c2pc5zmcsyjre", + "tokens": 7940, + "fixed_tokens": 1093, + "processed_tokens": 59768, + "tool_calls": 5, + "turns": 6 + }, + "render-012": { + "answer": "deferred", + "tokens": 9407, + "fixed_tokens": 1097, + "processed_tokens": 87507, + "tool_calls": 7, + "turns": 8 + }, + "render-013": { + "answer": "16-bit float per channel", + "tokens": 12974, + "fixed_tokens": 1096, + "processed_tokens": 81675, + "tool_calls": 6, + "turns": 7 + }, + "render-014": { + "answer": "52 percent", + "tokens": 8240, + "fixed_tokens": 1096, + "processed_tokens": 60473, + "tool_calls": 5, + "turns": 6 + }, + "render-015": { + "answer": "4 minutes", + "tokens": 7825, + "fixed_tokens": 1096, + "processed_tokens": 59022, + "tool_calls": 5, + "turns": 6 + }, + "render-016": { + "answer": "1.2 km", + "tokens": 8047, + "fixed_tokens": 1090, + "processed_tokens": 59792, + "tool_calls": 5, + "turns": 6 + }, + "render-017": { + "answer": "unknown", + "tokens": 13698, + "fixed_tokens": 1098, + "processed_tokens": 97915, + "tool_calls": 7, + "turns": 8 + }, + "render-018": { + "answer": "unknown", + "tokens": 12535, + "fixed_tokens": 1098, + "processed_tokens": 80034, + "tool_calls": 6, + "turns": 7 + }, + "render-019": { + "answer": "dec-use-the-custom-brine-curve-for-tonemapping--p735g6wf2njare86", + "tokens": 7462, + "fixed_tokens": 1090, + "processed_tokens": 57698, + "tool_calls": 5, + "turns": 6 + }, + "render-020": { + "answer": "fnd-screen-space-refraction-loses-off-screen-content--k8zvmmdee7wg2fjy", + "tokens": 7410, + "fixed_tokens": 1094, + "processed_tokens": 57706, + "tool_calls": 5, + "turns": 6 + }, + "render-021": { + "answer": "dec-adopt-a-visibility-buffer-for-all-opaque-geometr--68vchzqgd1avw8q4", + "tokens": 10677, + "fixed_tokens": 1096, + "processed_tokens": 89339, + "tool_calls": 7, + "turns": 8 + }, + "render-022": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 13591, + "fixed_tokens": 1089, + "processed_tokens": 114872, + "tool_calls": 8, + "turns": 9 + }, + "render-023": { + "answer": "dec-capture-the-cliffs-and-dunes-benchmarks-on-every--3pqyr66yh4v6q676", + "tokens": 7166, + "fixed_tokens": 1091, + "processed_tokens": 56604, + "tool_calls": 5, + "turns": 6 + }, + "render-024": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 7893, + "fixed_tokens": 1100, + "processed_tokens": 59463, + "tool_calls": 5, + "turns": 6 + }, + "render-025": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-after-fp16-pa--p4gcmzw2a4bvzjev", + "tokens": 18103, + "fixed_tokens": 1098, + "processed_tokens": 161381, + "tool_calls": 10, + "turns": 10 + }, + "render-026": { + "answer": "dec-ship-four-tern-quality-presets--vjw18d82dx5gath8", + "tokens": 7883, + "fixed_tokens": 1095, + "processed_tokens": 70139, + "tool_calls": 6, + "turns": 7 + }, + "render-027": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 10308, + "fixed_tokens": 1097, + "processed_tokens": 99224, + "tool_calls": 8, + "turns": 9 + }, + "render-028": { + "answer": "dec-move-to-gpu-driven-two-phase-occlusion-culling--6dre481mabxsw6ee", + "tokens": 7536, + "fixed_tokens": 1100, + "processed_tokens": 57815, + "tool_calls": 5, + "turns": 6 + }, + "render-029": { + "answer": "fnd-vsm-page-pool-overflows-during-spring-tides--m2p8e89hhdmh0z8a", + "tokens": 8284, + "fixed_tokens": 1099, + "processed_tokens": 60559, + "tool_calls": 5, + "turns": 6 + }, + "render-030": { + "answer": "dec-use-screen-space-refraction-for-saltglass--vknzksgdrgr7zkg6", + "tokens": 22395, + "fixed_tokens": 1099, + "processed_tokens": 130991, + "tool_calls": 9, + "turns": 8 + }, + "render-031": { + "answer": "Alpha-tested foliage was moved out of the visibility buffer because it cost 2.1 ms there versus 0.8 ms in forward rendering, so it stays in a forward pass after the resolve; saltglass was moved out because it refracts the resolved scene and so needs that scene to be rendered first (record dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9).", + "tokens": 8513, + "fixed_tokens": 1094, + "processed_tokens": 59627, + "tool_calls": 5, + "turns": 6 + }, + "render-032": { + "answer": "The team chose Tern because Halcyon has no Skiff build, and shipping it on two platforms would break the one-upscaler constraint; owning the code also mattered more for saltglass than raw quality, since the team can change Tern the same week without a vendor ticket or license (dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e).", + "tokens": 7956, + "fixed_tokens": 1097, + "processed_tokens": 58727, + "tool_calls": 5, + "turns": 6 + }, + "render-033": { + "answer": "128 MB", + "tokens": 8166, + "fixed_tokens": 1096, + "processed_tokens": 60310, + "tool_calls": 5, + "turns": 6 + }, + "render-034": { + "answer": "PC workstation (Meridian clock profile)", + "tokens": 10112, + "fixed_tokens": 1100, + "processed_tokens": 64788, + "tool_calls": 7, + "turns": 6 + }, + "render-035": { + "answer": "5912 pages", + "tokens": 9464, + "fixed_tokens": 1096, + "processed_tokens": 87617, + "tool_calls": 7, + "turns": 8 + }, + "render-036": { + "answer": "HDR-07", + "tokens": 7644, + "fixed_tokens": 1095, + "processed_tokens": 58160, + "tool_calls": 5, + "turns": 6 + }, + "render-037": { + "answer": "0.33", + "tokens": 11937, + "fixed_tokens": 1097, + "processed_tokens": 148313, + "tool_calls": 11, + "turns": 12 + }, + "render-038": { + "answer": "dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9", + "tokens": 17820, + "fixed_tokens": 1092, + "processed_tokens": 150639, + "tool_calls": 10, + "turns": 11 + }, + "render-039": { + "answer": "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "tokens": 8146, + "fixed_tokens": 1097, + "processed_tokens": 56380, + "tool_calls": 7, + "turns": 6 + }, + "render-040": { + "answer": "fnd-lantern-at-alpha--ckpd9ecvdjqbcr7c", + "tokens": 10554, + "fixed_tokens": 1096, + "processed_tokens": 119153, + "tool_calls": 10, + "turns": 11 + }, + "render-041": { + "answer": "4.7 ms", + "tokens": 8423, + "fixed_tokens": 1093, + "processed_tokens": 60572, + "tool_calls": 5, + "turns": 6 + }, + "render-042": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 8310, + "fixed_tokens": 1094, + "processed_tokens": 60383, + "tool_calls": 5, + "turns": 6 + }, + "render-043": { + "answer": "fnd-meshlet-cone-culling-rejects-9-percent-of-triang--t6e5x7trzedex105", + "tokens": 7210, + "fixed_tokens": 1087, + "processed_tokens": 56953, + "tool_calls": 5, + "turns": 6 + }, + "render-044": { + "answer": "dec-screen-space-refraction-with-a-ray-marched-fallb--2639xtpyrfrvbx1r", + "tokens": 7906, + "fixed_tokens": 1097, + "processed_tokens": 59056, + "tool_calls": 5, + "turns": 6 + }, + "render-045": { + "answer": false, + "tokens": 7670, + "fixed_tokens": 1092, + "processed_tokens": 58300, + "tool_calls": 5, + "turns": 6 + }, + "render-046": { + "answer": "The cost finding that sank virtual shadow maps (3.1 ms) was a capture error; with cached pages, VSM costs 1.4 ms, less than four cascades, so the sun decision was reopened (record dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4).", + "tokens": 8761, + "fixed_tokens": 1091, + "processed_tokens": 60844, + "tool_calls": 5, + "turns": 6 + }, + "render-047": { + "answer": "Because a release-build capture showed the visibility buffer on Skiff at 7.4 ms versus 8.1 ms for forward+ (6.3 ms with half-rate material resolve), so the forward+ fallback was no longer worth keeping, and maintaining two opaque paths cost about one engineer-week a month. Record: dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 8280, + "fixed_tokens": 1096, + "processed_tokens": 59567, + "tool_calls": 5, + "turns": 6 + }, + "render-048": { + "answer": "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "tokens": 26905, + "fixed_tokens": 1093, + "processed_tokens": 215642, + "tool_calls": 12, + "turns": 13 + }, + "render-049": { + "answer": "dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4", + "tokens": 9198, + "fixed_tokens": 1092, + "processed_tokens": 84980, + "tool_calls": 7, + "turns": 8 + }, + "render-050": { + "answer": "con-frame-budget--fzdrbxns6dmrrmy3", + "tokens": 6962, + "fixed_tokens": 1094, + "processed_tokens": 64226, + "tool_calls": 6, + "turns": 7 + }, + "render-051": { + "answer": "iss-tern-ghosting-on-moving-saltglass--2yr3q6zf476wajqp", + "tokens": 11283, + "fixed_tokens": 1092, + "processed_tokens": 95365, + "tool_calls": 9, + "turns": 8 + }, + "render-052": { + "answer": "61 percent", + "tokens": 7657, + "fixed_tokens": 1099, + "processed_tokens": 58302, + "tool_calls": 5, + "turns": 6 + }, + "render-053": { + "answer": "every_frame", + "tokens": 9380, + "fixed_tokens": 1100, + "processed_tokens": 86797, + "tool_calls": 7, + "turns": 8 + }, + "render-054": { + "answer": "debug-gpu-validation", + "tokens": 12728, + "fixed_tokens": 1099, + "processed_tokens": 155651, + "tool_calls": 11, + "turns": 12 + }, + "render-055": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-with-the-reac--81tcpzvhwzcy6wyn", + "tokens": 14213, + "fixed_tokens": 1094, + "processed_tokens": 134337, + "tool_calls": 11, + "turns": 10 + }, + "render-056": { + "answer": "208,000 meshlets", + "tokens": 11823, + "fixed_tokens": 1093, + "processed_tokens": 149391, + "tool_calls": 11, + "turns": 12 + }, + "render-057": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 16558, + "fixed_tokens": 1100, + "processed_tokens": 165343, + "tool_calls": 10, + "turns": 11 + }, + "render-058": { + "answer": "iss-which-upscaler-should-lantern-ship--g8c2y9tsr856tn2v", + "tokens": 17575, + "fixed_tokens": 1101, + "processed_tokens": 190137, + "tool_calls": 12, + "turns": 12 + }, + "render-059": { + "answer": "fnd-sdf-gi-costs-4-7-ms-on-meridian--sm19rt2f274ajsvr", + "tokens": 11112, + "fixed_tokens": 1100, + "processed_tokens": 119681, + "tool_calls": 9, + "turns": 10 + }, + "render-060": { + "answer": "0.31", + "tokens": 11898, + "fixed_tokens": 1095, + "processed_tokens": 148222, + "tool_calls": 11, + "turns": 12 + }, + "render-061": { + "answer": "con-ui-composites-at-a-fixed-203-nits-in-hdr--de6aft7cvhdezn5s", + "tokens": 9145, + "fixed_tokens": 1099, + "processed_tokens": 86022, + "tool_calls": 7, + "turns": 8 + }, + "render-062": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 9424, + "fixed_tokens": 1092, + "processed_tokens": 96156, + "tool_calls": 8, + "turns": 9 + }, + "render-063": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 13534, + "fixed_tokens": 1097, + "processed_tokens": 88143, + "tool_calls": 8, + "turns": 7 + }, + "render-064": { + "answer": "dec-use-a-cached-shadow-atlas-for-local-lights--8p6f2v774he38t5j", + "tokens": 10978, + "fixed_tokens": 1090, + "processed_tokens": 65947, + "tool_calls": 9, + "turns": 6 + }, + "render-065": { + "answer": "dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e", + "tokens": 9021, + "fixed_tokens": 1098, + "processed_tokens": 85697, + "tool_calls": 7, + "turns": 8 + }, + "render-066": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 9097, + "fixed_tokens": 1091, + "processed_tokens": 61499, + "tool_calls": 5, + "turns": 6 + }, + "render-067": { + "answer": 4, + "tokens": 9707, + "fixed_tokens": 1091, + "processed_tokens": 100816, + "tool_calls": 9, + "turns": 10 + }, + "render-068": { + "answer": [ + "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "dec-lock-caustic-photon-jitter-to-the-tide-phase--85x10k4h11t8nx7c", + "dec-use-the-probe-grid-alone-in-120-hz-performance-m--6dzm1nt40pjfpyae", + "dec-trace-sdf-gi-at-quarter-resolution-in-performanc--06cqt3mqn3fg5h6e", + "dec-read-paper-white-from-the-meridian-system-settin--7crbx73gqw1hn7xq", + "dec-add-an-in-game-paper-white-slider-from-80-to-400--1phfvxby983z8tjf" + ], + "tokens": 27429, + "fixed_tokens": 1101, + "processed_tokens": 220851, + "tool_calls": 27, + "turns": 12 + }, + "render-069": { + "answer": "iss-shadow-pages-run-out-at-spring-low-tide--f4fgy6b44jbkrsmh", + "tokens": 17121, + "fixed_tokens": 1094, + "processed_tokens": 133934, + "tool_calls": 10, + "turns": 11 + }, + "render-070": { + "answer": [ + "dec-use-signed-distance-field-gi--yn366zmp3k7snd9w", + "dec-feed-a-saltglass-reactive-mask-into-tern--xezjs7v3vq3r8e6k" + ], + "tokens": 70582, + "fixed_tokens": 1090, + "processed_tokens": 959599, + "tool_calls": 23, + "turns": 24 + }, + "render-071": { + "answer": 3, + "tokens": 9480, + "fixed_tokens": 1096, + "processed_tokens": 56869, + "tool_calls": 5, + "turns": 6 + }, + "render-072": { + "answer": [ + "rsk-sdf-volumes-outgrow-the-render-target-memory-bud--xfd2pfztvskdcd2c", + "rsk-artists-cannot-match-concept-paintings-under-the--6cfh9vkamn0kbtcx" + ], + "tokens": 12129, + "fixed_tokens": 1098, + "processed_tokens": 90545, + "tool_calls": 9, + "turns": 8 + }, + "render-073": { + "answer": [ + "rsk-vsm-page-pool-overflows-when-tides-expose-the-fl--x1r161nrdngra4ty", + "rsk-gi-cannot-fit-the-120-hz-performance-mode--tqt3fj839dqfq5s7", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 11781, + "fixed_tokens": 1096, + "processed_tokens": 93783, + "tool_calls": 10, + "turns": 9 + }, + "render-074": { + "answer": 5, + "tokens": 7590, + "fixed_tokens": 1097, + "processed_tokens": 53633, + "tool_calls": 5, + "turns": 6 + }, + "render-075": { + "answer": 3, + "tokens": 43395, + "fixed_tokens": 1089, + "processed_tokens": 1143497, + "tool_calls": 36, + "turns": 37 + }, + "render-076": { + "answer": false, + "tokens": 9844, + "fixed_tokens": 1102, + "processed_tokens": 86217, + "tool_calls": 7, + "turns": 8 + }, + "render-077": { + "answer": true, + "tokens": 19694, + "fixed_tokens": 1100, + "processed_tokens": 264408, + "tool_calls": 16, + "turns": 17 + }, + "render-078": { + "answer": [ + "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 20102, + "fixed_tokens": 1100, + "processed_tokens": 238183, + "tool_calls": 14, + "turns": 15 + }, + "render-079": { + "answer": 5, + "tokens": 45854, + "fixed_tokens": 1092, + "processed_tokens": 809729, + "tool_calls": 30, + "turns": 27 + }, + "render-080": { + "answer": "6 versions", + "tokens": 15540, + "fixed_tokens": 1096, + "processed_tokens": 165339, + "tool_calls": 11, + "turns": 12 + }, + "render-081": { + "answer": "unknown", + "tokens": 16007, + "fixed_tokens": 1096, + "processed_tokens": 118843, + "tool_calls": 10, + "turns": 9 + }, + "render-082": { + "answer": false, + "tokens": 14727, + "fixed_tokens": 1100, + "processed_tokens": 166456, + "tool_calls": 11, + "turns": 12 + }, + "render-083": { + "answer": 0, + "tokens": 33198, + "fixed_tokens": 1097, + "processed_tokens": 373512, + "tool_calls": 16, + "turns": 17 + }, + "tools-001": { + "answer": "completed", + "tokens": 9117, + "fixed_tokens": 1094, + "processed_tokens": 61725, + "tool_calls": 5, + "turns": 6 + }, + "tools-002": { + "answer": "47 minutes", + "tokens": 7671, + "fixed_tokens": 1095, + "processed_tokens": 58194, + "tool_calls": 5, + "turns": 6 + }, + "tools-003": { + "answer": "6 min 40 s", + "tokens": 8132, + "fixed_tokens": 1098, + "processed_tokens": 59919, + "tool_calls": 5, + "turns": 6 + }, + "tools-004": { + "answer": "unknown", + "tokens": 18323, + "fixed_tokens": 1093, + "processed_tokens": 111690, + "tool_calls": 7, + "turns": 8 + }, + "tools-005": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 8337, + "fixed_tokens": 1097, + "processed_tokens": 59858, + "tool_calls": 5, + "turns": 6 + }, + "tools-006": { + "answer": "dec-build-a-content-addressed-shader-compile-cache-s--e65rq9439k7xz5zh", + "tokens": 9632, + "fixed_tokens": 1096, + "processed_tokens": 87005, + "tool_calls": 7, + "turns": 8 + }, + "tools-007": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 33711, + "fixed_tokens": 1099, + "processed_tokens": 433004, + "tool_calls": 20, + "turns": 21 + }, + "tools-008": { + "answer": "dec-key-the-shader-cache-on-raw-source-text-and-comp--hp6awmx3gq5rn9nc", + "tokens": 9478, + "fixed_tokens": 1094, + "processed_tokens": 86303, + "tool_calls": 7, + "turns": 8 + }, + "tools-009": { + "answer": "92.4%", + "tokens": 7452, + "fixed_tokens": 1096, + "processed_tokens": 57661, + "tool_calls": 5, + "turns": 6 + }, + "tools-010": { + "answer": "dec-cap-each-farm-agent-at-6-parallel-shader-compile--q5w4sfjm73tsj8x8", + "tokens": 14615, + "fixed_tokens": 1094, + "processed_tokens": 113235, + "tool_calls": 8, + "turns": 9 + }, + "tools-011": { + "answer": "fnd-two-compiler-builds-with-the-same-version-string--gmgr4ha0ndxqmcn9", + "tokens": 12896, + "fixed_tokens": 1100, + "processed_tokens": 127452, + "tool_calls": 9, + "turns": 10 + }, + "tools-012": { + "answer": 3, + "tokens": 20826, + "fixed_tokens": 1096, + "processed_tokens": 173161, + "tool_calls": 9, + "turns": 10 + }, + "tools-013": { + "answer": 8, + "tokens": 114125, + "fixed_tokens": 1104, + "processed_tokens": 2921245, + "tool_calls": 43, + "turns": 44 + }, + "tools-014": { + "answer": false, + "tokens": 12635, + "fixed_tokens": 1100, + "processed_tokens": 104812, + "tool_calls": 8, + "turns": 9 + }, + "tools-015": { + "answer": [ + "iss-whether-the-artist-sandbox-should-load-on-a-work--73j819anqct8etfn", + "iss-whether-the-audio-bank-packer-should-load-on-a-w--sqzp2rwp6s01w3b0", + "iss-whether-the-level-editor-autosave-needs-its-own--rkvmnbqe7w644zkr", + "iss-whether-the-review-bot-should-load-on-a-worker-t--rtsjwvsksjdq15v1", + "iss-whether-the-level-editor-autosave-should-be-cach--radcjs4hyvh7p55q", + "iss-whether-the-crash-symbol-server-should-load-on-a--qvyphhgnsnm2cvqb", + "iss-whether-the-license-scanner-should-be-cached-bet--nqhd89czbfat9rya", + "iss-whether-the-doc-site-should-load-on-a-worker-thr--qxswe0pg52ekky0z", + "iss-whether-the-asset-search-should-load-on-a-worker--f9gpv3z522x6p6h4", + "iss-whether-the-changelog-bot-should-load-on-a-worke--jgtdn2f02p8fx5nr", + "iss-whether-the-review-bot-needs-its-own-memory-pool--pg3m8g9gn3y6v3kq", + "iss-whether-the-changelog-bot-should-skip-work-when--p2vrp5d682p3jrc3", + "iss-whether-the-patch-differ-should-load-on-a-worker--cysv4n56q1qn7t2k", + "iss-whether-the-artist-sandbox-should-skip-work-when--11tqzws72gms2h8x", + "iss-whether-the-audio-bank-packer-needs-its-own-memo--qjrqrseast8d5mhf", + "iss-whether-the-changelog-bot-needs-its-own-memory-p--mw8zejz6hk5m60x0", + "iss-whether-the-crash-symbol-server-should-ship-in-t--bhq91ajavah18wsk", + "iss-whether-the-asset-search-should-load-on-a-worker--80b27kqh36rpafdt", + "iss-whether-the-doc-site-should-ship-in-the-first-pa--wcdkq5zab7dxe5yj", + "iss-whether-the-patch-differ-should-be-cached-betwee--28b5zqb1ketwzf0r", + "iss-whether-the-test-runner-should-load-on-a-worker--aygay1r0qj4mzf78", + "iss-whether-the-texture-viewer-should-be-split-into--1jhgyqb70h5x7m7b", + "iss-whether-the-level-editor-autosave-should-skip-wo--fyysrcv6fvbsvcsy", + "iss-whether-the-level-editor-autosave-should-be-spli--ncnwh0csw671fs0q", + "iss-whether-the-asset-search-should-ship-in-the-firs--hsb2avhx613prxew", + "iss-whether-the-test-runner-needs-its-own-memory-poo--bs6arq0qv37d6wmx", + "iss-whether-the-artist-sandbox-should-be-split-into--ek47ca2vgmrjanp8", + "rsk-shader-cache-storage-could-outgrow-the-4-tb-farm--76y9649mwrx7aavd" + ], + "tokens": 67952, + "fixed_tokens": 1101, + "processed_tokens": 708431, + "tool_calls": 19, + "turns": 19 + }, + "tools-016": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 12770, + "fixed_tokens": 1100, + "processed_tokens": 112412, + "tool_calls": 8, + "turns": 9 + }, + "world-001": { + "answer": "64 KB", + "tokens": 7869, + "fixed_tokens": 1092, + "processed_tokens": 59102, + "tool_calls": 5, + "turns": 6 + }, + "world-002": { + "answer": "5.1 ms", + "tokens": 8086, + "fixed_tokens": 1090, + "processed_tokens": 59632, + "tool_calls": 5, + "turns": 6 + }, + "world-003": { + "answer": "1 crash per 900 script waits", + "tokens": 14037, + "fixed_tokens": 1097, + "processed_tokens": 93895, + "tool_calls": 6, + "turns": 7 + }, + "world-004": { + "answer": "4,100 file opens", + "tokens": 11607, + "fixed_tokens": 1092, + "processed_tokens": 78889, + "tool_calls": 6, + "turns": 7 + }, + "world-005": { + "answer": "11 m", + "tokens": 8106, + "fixed_tokens": 1098, + "processed_tokens": 59574, + "tool_calls": 5, + "turns": 6 + }, + "world-006": { + "answer": "3.4 times per minute", + "tokens": 8035, + "fixed_tokens": 1095, + "processed_tokens": 59593, + "tool_calls": 5, + "turns": 6 + }, + "world-007": { + "answer": "260 MB", + "tokens": 9440, + "fixed_tokens": 1092, + "processed_tokens": 85777, + "tool_calls": 7, + "turns": 8 + }, + "world-008": { + "answer": "10 Hz", + "tokens": 7815, + "fixed_tokens": 1101, + "processed_tokens": 59072, + "tool_calls": 5, + "turns": 6 + }, + "world-009": { + "answer": "0.3 ms", + "tokens": 7538, + "fixed_tokens": 1090, + "processed_tokens": 57933, + "tool_calls": 5, + "turns": 6 + }, + "world-010": { + "answer": "unknown", + "tokens": 6948, + "fixed_tokens": 1095, + "processed_tokens": 61808, + "tool_calls": 6, + "turns": 7 + }, + "world-011": { + "answer": "fnd-shipping-entity-names-costs-57-mb-at-full-popula--6w7gjvy7wdmdx9qx", + "tokens": 12155, + "fixed_tokens": 1088, + "processed_tokens": 80932, + "tool_calls": 8, + "turns": 7 + }, + "world-012": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 7782, + "fixed_tokens": 1094, + "processed_tokens": 58711, + "tool_calls": 5, + "turns": 6 + }, + "world-013": { + "answer": "unknown", + "tokens": 10894, + "fixed_tokens": 1091, + "processed_tokens": 76944, + "tool_calls": 6, + "turns": 7 + }, + "world-014": { + "answer": "unknown", + "tokens": 27171, + "fixed_tokens": 1090, + "processed_tokens": 182252, + "tool_calls": 9, + "turns": 10 + }, + "world-015": { + "answer": "3.0 ms", + "tokens": 8418, + "fixed_tokens": 1097, + "processed_tokens": 70040, + "tool_calls": 6, + "turns": 7 + }, + "world-016": { + "answer": "unknown", + "tokens": 6928, + "fixed_tokens": 1090, + "processed_tokens": 64345, + "tool_calls": 6, + "turns": 7 + }, + "world-017": { + "answer": "dec-partition-the-world-into-128-m-cells--372b6czbjf29ymaw", + "tokens": 13026, + "fixed_tokens": 1101, + "processed_tokens": 107263, + "tool_calls": 8, + "turns": 9 + }, + "world-018": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 11456, + "fixed_tokens": 1098, + "processed_tokens": 106927, + "tool_calls": 9, + "turns": 9 + }, + "world-019": { + "answer": "fnd-tide-priority-reads-starve-player-proximity-load--by0d7b0bk39xz3rk", + "tokens": 7863, + "fixed_tokens": 1095, + "processed_tokens": 59007, + "tool_calls": 5, + "turns": 6 + }, + "world-020": { + "answer": "dec-schedule-streaming-reads-by-deadline-instead-of--q2x5aknabgcssdj2", + "tokens": 9008, + "fixed_tokens": 1093, + "processed_tokens": 85023, + "tool_calls": 7, + "turns": 8 + }, + "world-021": { + "answer": "dec-cut-navigation-tiles-to-64-m-so-three-fit-along--snvepk8ygeb3qsjs", + "tokens": 7530, + "fixed_tokens": 1091, + "processed_tokens": 58275, + "tool_calls": 5, + "turns": 6 + }, + "world-022": { + "answer": "fnd-41-of-alpha-crash-reports-have-no-usable-stack--35a50b07ferhf13m", + "tokens": 7434, + "fixed_tokens": 1091, + "processed_tokens": 57695, + "tool_calls": 5, + "turns": 6 + }, + "world-023": { + "answer": "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "tokens": 9778, + "fixed_tokens": 1088, + "processed_tokens": 87970, + "tool_calls": 7, + "turns": 8 + }, + "world-024": { + "answer": "dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew", + "tokens": 9408, + "fixed_tokens": 1092, + "processed_tokens": 86584, + "tool_calls": 7, + "turns": 8 + }, + "world-025": { + "answer": "con-streaming-pool-budget--kca88sgh2j6sxvd5", + "tokens": 9085, + "fixed_tokens": 1094, + "processed_tokens": 86253, + "tool_calls": 7, + "turns": 8 + }, + "world-026": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 7738, + "fixed_tokens": 1093, + "processed_tokens": 58916, + "tool_calls": 5, + "turns": 6 + }, + "world-027": { + "answer": "160 MB/s", + "tokens": 7816, + "fixed_tokens": 1093, + "processed_tokens": 58874, + "tool_calls": 5, + "turns": 6 + }, + "world-028": { + "answer": "fnd-streaming-tide-heights-costs-46-kb-s-per-peer--knxqtyqh6xcz1tp9", + "tokens": 7562, + "fixed_tokens": 1094, + "processed_tokens": 58271, + "tool_calls": 5, + "turns": 6 + }, + "world-029": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 10315, + "fixed_tokens": 1089, + "processed_tokens": 88059, + "tool_calls": 7, + "turns": 8 + }, + "world-030": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 11289, + "fixed_tokens": 1089, + "processed_tokens": 90000, + "tool_calls": 9, + "turns": 8 + }, + "world-031": { + "answer": "1.5 GB/s", + "tokens": 8420, + "fixed_tokens": 1096, + "processed_tokens": 61257, + "tool_calls": 5, + "turns": 6 + }, + "world-032": { + "answer": "fnd-float-tide-heights-drift-apart-between-pc-and-me--7jb0fg4b5xzr7q5e", + "tokens": 9973, + "fixed_tokens": 1091, + "processed_tokens": 88539, + "tool_calls": 7, + "turns": 8 + }, + "world-033": { + "answer": "dec-save-player-changes-as-deltas-against-the-tide-f--n38scy1yaarsr8cw", + "tokens": 26051, + "fixed_tokens": 1094, + "processed_tokens": 221563, + "tool_calls": 12, + "turns": 11 + }, + "world-034": { + "answer": "unknown", + "tokens": 12168, + "fixed_tokens": 1087, + "processed_tokens": 93531, + "tool_calls": 7, + "turns": 8 + }, + "world-035": { + "answer": false, + "tokens": 7988, + "fixed_tokens": 1094, + "processed_tokens": 58965, + "tool_calls": 5, + "turns": 6 + }, + "world-036": { + "answer": "Kestrel retired fibers for engine jobs because 41% of alpha crash reports had no usable stack (fiber stacks made crash dumps unreadable), which blocked crash triage for beta, and fibers gained only about two points of utilisation (95% vs 93% for the continuation prototype) while also causing the Gull thread-local bug. Record: dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h", + "tokens": 8372, + "fixed_tokens": 1101, + "processed_tokens": 59592, + "tool_calls": 5, + "turns": 6 + }, + "world-037": { + "answer": "fnd-hot-reload-drops-suspended-gull-coroutines--va0fn0svryzwbzws", + "tokens": 7951, + "fixed_tokens": 1096, + "processed_tokens": 58810, + "tool_calls": 5, + "turns": 6 + }, + "world-038": { + "answer": "3 archetypes", + "tokens": 12470, + "fixed_tokens": 1099, + "processed_tokens": 139506, + "tool_calls": 11, + "turns": 11 + }, + "world-039": { + "answer": "1.2 GB/s", + "tokens": 8051, + "fixed_tokens": 1097, + "processed_tokens": 59678, + "tool_calls": 5, + "turns": 6 + }, + "world-040": { + "answer": "unknown", + "tokens": 8146, + "fixed_tokens": 1091, + "processed_tokens": 60387, + "tool_calls": 5, + "turns": 6 + }, + "world-041": { + "answer": "61%", + "tokens": 7814, + "fixed_tokens": 1093, + "processed_tokens": 59039, + "tool_calls": 5, + "turns": 6 + }, + "world-042": { + "answer": "5.4 s", + "tokens": 9240, + "fixed_tokens": 1098, + "processed_tokens": 86563, + "tool_calls": 7, + "turns": 8 + }, + "world-043": { + "answer": "5.8 min", + "tokens": 10545, + "fixed_tokens": 1095, + "processed_tokens": 115698, + "tool_calls": 9, + "turns": 10 + }, + "world-044": { + "answer": "dec-move-engine-jobs-to-fibers--qjkgnm8cex8ydzfq", + "tokens": 23067, + "fixed_tokens": 1100, + "processed_tokens": 274707, + "tool_calls": 15, + "turns": 14 + }, + "world-045": { + "answer": "dec-run-engine-jobs-on-fibers-and-pin-gull-scripts-t--rw28cbqfd07rd9te", + "tokens": 19052, + "fixed_tokens": 1098, + "processed_tokens": 182065, + "tool_calls": 11, + "turns": 12 + }, + "world-046": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 9940, + "fixed_tokens": 1092, + "processed_tokens": 89580, + "tool_calls": 7, + "turns": 8 + }, + "world-047": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 10980, + "fixed_tokens": 1095, + "processed_tokens": 98766, + "tool_calls": 8, + "turns": 9 + }, + "world-048": { + "answer": "dec-compile-gull-to-bytecode-for-a-register-vm--jrqje646fpevp5f3", + "tokens": 7852, + "fixed_tokens": 1088, + "processed_tokens": 59186, + "tool_calls": 5, + "turns": 6 + }, + "world-049": { + "answer": "22%", + "tokens": 8008, + "fixed_tokens": 1090, + "processed_tokens": 59557, + "tool_calls": 5, + "turns": 6 + }, + "world-050": { + "answer": "fnd-alpha-content-needs-6-2-ms-of-gull-time-per-fram--4mekgdpsdv5nk085", + "tokens": 12540, + "fixed_tokens": 1092, + "processed_tokens": 105899, + "tool_calls": 8, + "turns": 9 + }, + "world-051": { + "answer": "Float tides drifted apart between machines, and strict IEEE flags fixed only the compiler, not Skiff's math library, whose sin() differs by one ulp (fnd-strict-float-mode-still-diverges-because-skiff-s--z811r28v0a1wzb13). Fixed-point 32.32 arithmetic uses only integer adds and multiplies plus a build-time sine table, so it gives identical bits on every CPU and compiler (dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e).", + "tokens": 9714, + "fixed_tokens": 1097, + "processed_tokens": 86411, + "tool_calls": 7, + "turns": 8 + }, + "world-052": { + "answer": "Dara Whitlock", + "tokens": 8070, + "fixed_tokens": 1094, + "processed_tokens": 59889, + "tool_calls": 5, + "turns": 6 + }, + "world-053": { + "answer": "1.9 ms", + "tokens": 9587, + "fixed_tokens": 1098, + "processed_tokens": 86411, + "tool_calls": 7, + "turns": 8 + }, + "world-054": { + "answer": "Marta Lindqvist", + "tokens": 13488, + "fixed_tokens": 1098, + "processed_tokens": 154875, + "tool_calls": 11, + "turns": 12 + }, + "world-055": { + "answer": "Kelp stayed because the team shrank its dictionary to 1 MB, which beat Zstandard cold on Skiff (2.2 GB/s vs 1.3 GB/s, per the superseded finding), so the 4 MB dictionary's loss was fixed rather than abandoning Kelp; see dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew. Firmware 3.2 later cut the margin to 1.5 vs 1.2 GB/s, but Kelp still wins.", + "tokens": 10381, + "fixed_tokens": 1093, + "processed_tokens": 64739, + "tool_calls": 7, + "turns": 6 + }, + "world-056": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 11435, + "fixed_tokens": 1099, + "processed_tokens": 93810, + "tool_calls": 9, + "turns": 8 + }, + "world-057": { + "answer": "(212, 87)", + "tokens": 12565, + "fixed_tokens": 1096, + "processed_tokens": 180967, + "tool_calls": 13, + "turns": 14 + }, + "world-058": { + "answer": "0.8 GB/s", + "tokens": 10440, + "fixed_tokens": 1094, + "processed_tokens": 91650, + "tool_calls": 7, + "turns": 8 + }, + "world-059": { + "answer": "dec-derive-the-missing-harmonic-phase-from-each-save--gbqrvvgyq2gy2gnw", + "tokens": 12798, + "fixed_tokens": 1093, + "processed_tokens": 127808, + "tool_calls": 11, + "turns": 10 + }, + "world-060": { + "answer": "dec-partition-the-world-into-192-m-cells--bk9r1c5jwb6fh8tc", + "tokens": 9523, + "fixed_tokens": 1098, + "processed_tokens": 86673, + "tool_calls": 7, + "turns": 8 + }, + "world-061": { + "answer": "236 MB", + "tokens": 13741, + "fixed_tokens": 1091, + "processed_tokens": 187378, + "tool_calls": 13, + "turns": 14 + }, + "world-062": { + "answer": "310 MB", + "tokens": 14521, + "fixed_tokens": 1093, + "processed_tokens": 192764, + "tool_calls": 13, + "turns": 14 + }, + "world-063": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 15050, + "fixed_tokens": 1094, + "processed_tokens": 203828, + "tool_calls": 14, + "turns": 15 + }, + "world-064": { + "answer": [ + "dec-force-strict-ieee-float-mode-for-the-tide-simula--ecwmqt7ncjngysqw", + "dec-have-the-host-stream-tide-heights-to-every-peer--a3rb1seedn2bj7k8" + ], + "tokens": 21046, + "fixed_tokens": 1098, + "processed_tokens": 192243, + "tool_calls": 13, + "turns": 13 + }, + "world-065": { + "answer": "3.1 ms", + "tokens": 12845, + "fixed_tokens": 1090, + "processed_tokens": 154652, + "tool_calls": 11, + "turns": 12 + }, + "world-066": { + "answer": "46", + "tokens": 13036, + "fixed_tokens": 1097, + "processed_tokens": 114591, + "tool_calls": 8, + "turns": 9 + }, + "world-067": { + "answer": "1.2 GB/s", + "tokens": 10386, + "fixed_tokens": 1093, + "processed_tokens": 91357, + "tool_calls": 7, + "turns": 8 + }, + "world-068": { + "answer": true, + "tokens": 14363, + "fixed_tokens": 1096, + "processed_tokens": 141686, + "tool_calls": 11, + "turns": 12 + }, + "world-069": { + "answer": 6, + "tokens": 61727, + "fixed_tokens": 1096, + "processed_tokens": 746023, + "tool_calls": 20, + "turns": 21 + }, + "world-070": { + "answer": 3, + "tokens": 17648, + "fixed_tokens": 1094, + "processed_tokens": 153199, + "tool_calls": 13, + "turns": 12 + }, + "world-071": { + "answer": [ + "dec-hot-reload-gull-functions-by-swapping-bytecode-a--8dxnkg6h8vmcsr42", + "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "dec-evict-cells-by-distance-and-tide-forecast--fyp1jp366pjjzxcg", + "dec-store-components-in-archetype-tables-and-volatil--yxsjd26wcaz6fgx9" + ], + "tokens": 87258, + "fixed_tokens": 1093, + "processed_tokens": 1282510, + "tool_calls": 29, + "turns": 29 + }, + "world-072": { + "answer": [ + "dec-compress-streamed-assets-with-lz4--v1pfsgfp0b3yjv4s", + "dec-compress-streamed-assets-with-zstandard--bs1t15cygc51cpje" + ], + "tokens": 26173, + "fixed_tokens": 1090, + "processed_tokens": 242494, + "tool_calls": 14, + "turns": 15 + }, + "world-073": { + "answer": false, + "tokens": 10759, + "fixed_tokens": 1094, + "processed_tokens": 64412, + "tool_calls": 9, + "turns": 6 + }, + "world-074": { + "answer": 7, + "tokens": 9295, + "fixed_tokens": 1093, + "processed_tokens": 70360, + "tool_calls": 6, + "turns": 7 + }, + "world-075": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--8679sjn071ycsbpv", + "dec-partition-the-world-into-192-meter-cells--q62ddwdv2n8wxay6", + "dec-return-engine-jobs-to-the-thread-pool-with-conti--547qbweqvrp1j5g3" + ], + "tokens": 6325, + "fixed_tokens": 1094, + "processed_tokens": 50899, + "tool_calls": 5, + "turns": 6 + }, + "world-076": { + "answer": "unknown", + "tokens": 81916, + "fixed_tokens": 1093, + "processed_tokens": 2516521, + "tool_calls": 91, + "turns": 55 + }, + "world-077": { + "answer": 251, + "tokens": 20783, + "fixed_tokens": 1091, + "processed_tokens": 199095, + "tool_calls": 11, + "turns": 12 + }, + "world-078": { + "answer": [ + "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "rsk-reworking-the-read-scheduler-may-slip-the-coast--9jzvs7etvb5jxqtp" + ], + "tokens": 10347, + "fixed_tokens": 1086, + "processed_tokens": 58526, + "tool_calls": 5, + "turns": 6 + }, + "world-079": { + "answer": [ + "con-streaming-pool-budget--8czwb4fjr5p5vpy7", + "con-saves-from-beta-onward-must-load-in-every-later--v6z856sxtgc5sb09", + "con-no-runtime-code-generation-on-consoles--6cf25zqmsvqkdd73", + "con-tide-state-must-be-bit-identical-on-every-peer--zrpnp8qjyxzkas7t" + ], + "tokens": 13178, + "fixed_tokens": 1094, + "processed_tokens": 101995, + "tool_calls": 8, + "turns": 9 + }, + "world-080": { + "answer": [ + "iss-cross-platform-cloud-saves-are-unplanned--5bk0k4wrgj42ztfy", + "iss-gull-hot-reload-drops-suspended-coroutines--z7j2gmyny6gdxd07", + "iss-gull-hot-reload-rejects-struct-layout-changes--3v2w2f9hg2gwy6aj" + ], + "tokens": 27271, + "fixed_tokens": 1097, + "processed_tokens": 226999, + "tool_calls": 16, + "turns": 13 + }, + "world-081": { + "answer": "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "tokens": 34597, + "fixed_tokens": 1097, + "processed_tokens": 507098, + "tool_calls": 21, + "turns": 20 + } + } +} diff --git a/benchmarks/proof-recall/runs/x10-find-r1-nobrief.json b/benchmarks/proof-recall/runs/x10-find-r1-nobrief.json new file mode 100644 index 00000000..2538c389 --- /dev/null +++ b/benchmarks/proof-recall/runs/x10-find-r1-nobrief.json @@ -0,0 +1,2256 @@ +{ + "run": { + "name": "x10-find-r1-nobrief", + "protocol": "one fresh session per question", + "models": ["claude-haiku-5-5"], + "find_hash": "e5be653299abe9b2329822d98ffe09e90ee7cfe96f8711c813a51dc6500b887f" + }, + "answers": { + "ambiguous-001": { + "answer": "unknown", + "tokens": 48672, + "fixed_tokens": 1116, + "processed_tokens": 487436, + "tool_calls": 15, + "turns": 16 + }, + "ambiguous-002": { + "answer": "unknown", + "tokens": 44799, + "fixed_tokens": 1112, + "processed_tokens": 306641, + "tool_calls": 11, + "turns": 12 + }, + "ambiguous-003": { + "answer": "92.4%", + "tokens": 44016, + "fixed_tokens": 1117, + "processed_tokens": 882431, + "tool_calls": 27, + "turns": 27 + }, + "ambiguous-004": { + "answer": "6 min 40 s", + "tokens": 32490, + "fixed_tokens": 1119, + "processed_tokens": 529474, + "tool_calls": 20, + "turns": 21 + }, + "ambiguous-005": { + "answer": "unknown", + "tokens": 61801, + "fixed_tokens": 1109, + "processed_tokens": 1616894, + "tool_calls": 38, + "turns": 39 + }, + "ambiguous-006": { + "answer": "unknown", + "tokens": 71931, + "fixed_tokens": 1114, + "processed_tokens": 2021855, + "tool_calls": 45, + "turns": 46 + }, + "ambiguous-007": { + "answer": "32 steps", + "tokens": 25961, + "fixed_tokens": 1110, + "processed_tokens": 333008, + "tool_calls": 15, + "turns": 16 + }, + "ambiguous-008": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 19057, + "fixed_tokens": 1116, + "processed_tokens": 119610, + "tool_calls": 7, + "turns": 8 + }, + "ambiguous-009": { + "answer": "3.0 ms", + "tokens": 41336, + "fixed_tokens": 1116, + "processed_tokens": 393950, + "tool_calls": 17, + "turns": 15 + }, + "ambiguous-010": { + "answer": "160 MB/s", + "tokens": 25780, + "fixed_tokens": 1112, + "processed_tokens": 203971, + "tool_calls": 10, + "turns": 11 + }, + "ambiguous-011": { + "answer": "260 MB", + "tokens": 44256, + "fixed_tokens": 1111, + "processed_tokens": 350539, + "tool_calls": 12, + "turns": 13 + }, + "ambiguous-012": { + "answer": "11 m", + "tokens": 35647, + "fixed_tokens": 1113, + "processed_tokens": 469945, + "tool_calls": 17, + "turns": 18 + }, + "ambiguous-013": { + "answer": "61%", + "tokens": 34494, + "fixed_tokens": 1112, + "processed_tokens": 366541, + "tool_calls": 14, + "turns": 15 + }, + "ambiguous-014": { + "answer": "textures", + "tokens": 43888, + "fixed_tokens": 1112, + "processed_tokens": 768714, + "tool_calls": 23, + "turns": 24 + }, + "ambiguous-015": { + "answer": "unknown", + "tokens": 49680, + "fixed_tokens": 1114, + "processed_tokens": 570684, + "tool_calls": 16, + "turns": 17 + }, + "ambiguous-016": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 41111, + "fixed_tokens": 1115, + "processed_tokens": 877298, + "tool_calls": 28, + "turns": 29 + }, + "ambiguous-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 42676, + "fixed_tokens": 1111, + "processed_tokens": 935018, + "tool_calls": 31, + "turns": 30 + }, + "ambiguous-018": { + "answer": "unknown", + "tokens": 72118, + "fixed_tokens": 1111, + "processed_tokens": 3002238, + "tool_calls": 62, + "turns": 61 + }, + "ambiguous-019": { + "answer": "7.4 ms", + "tokens": 27392, + "fixed_tokens": 1116, + "processed_tokens": 228643, + "tool_calls": 11, + "turns": 12 + }, + "ambiguous-020": { + "answer": "5912 pages", + "tokens": 27667, + "fixed_tokens": 1112, + "processed_tokens": 484368, + "tool_calls": 22, + "turns": 23 + }, + "ambiguous-021": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 31250, + "fixed_tokens": 1112, + "processed_tokens": 364567, + "tool_calls": 15, + "turns": 16 + }, + "ambiguous-022": { + "answer": null, + "tokens": 109754, + "fixed_tokens": 1110, + "processed_tokens": 5969490, + "tool_calls": 88, + "turns": 86 + }, + "ambiguous-023": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 10647, + "fixed_tokens": 1118, + "processed_tokens": 96138, + "tool_calls": 8, + "turns": 9 + }, + "ambiguous-024": { + "answer": "unknown", + "tokens": 41302, + "fixed_tokens": 1114, + "processed_tokens": 821167, + "tool_calls": 27, + "turns": 28 + }, + "cross-001": { + "answer": "con-frame-budget--br0qszfg8fys908n", + "tokens": 24213, + "fixed_tokens": 1102, + "processed_tokens": 254772, + "tool_calls": 12, + "turns": 13 + }, + "cross-002": { + "answer": 2, + "tokens": 25850, + "fixed_tokens": 1105, + "processed_tokens": 204397, + "tool_calls": 14, + "turns": 12 + }, + "cross-003": { + "answer": 2, + "tokens": 19246, + "fixed_tokens": 1108, + "processed_tokens": 193953, + "tool_calls": 13, + "turns": 12 + }, + "cross-004": { + "answer": "dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e", + "tokens": 36378, + "fixed_tokens": 1104, + "processed_tokens": 384290, + "tool_calls": 14, + "turns": 15 + }, + "cross-005": { + "answer": "Rhea Valdés", + "tokens": 15375, + "fixed_tokens": 1105, + "processed_tokens": 227353, + "tool_calls": 16, + "turns": 17 + }, + "cross-006": { + "answer": "19%", + "tokens": 32709, + "fixed_tokens": 1103, + "processed_tokens": 496813, + "tool_calls": 19, + "turns": 20 + }, + "cross-007": { + "answer": "8 GB", + "tokens": 36042, + "fixed_tokens": 1105, + "processed_tokens": 254956, + "tool_calls": 11, + "turns": 12 + }, + "cross-008": { + "answer": "unknown", + "tokens": 31477, + "fixed_tokens": 1105, + "processed_tokens": 341517, + "tool_calls": 14, + "turns": 15 + }, + "cross-009": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 32699, + "fixed_tokens": 1109, + "processed_tokens": 284043, + "tool_calls": 14, + "turns": 13 + }, + "cross-010": { + "answer": "dec-build-an-on-device-pipeline-cache-for-skiff--726r9mtk76qndeh7", + "tokens": 24239, + "fixed_tokens": 1105, + "processed_tokens": 217229, + "tool_calls": 11, + "turns": 12 + }, + "cross-011": { + "answer": true, + "tokens": 47164, + "fixed_tokens": 1108, + "processed_tokens": 942804, + "tool_calls": 27, + "turns": 28 + }, + "cross-012": { + "answer": "unknown", + "tokens": 61045, + "fixed_tokens": 1103, + "processed_tokens": 2737310, + "tool_calls": 64, + "turns": 65 + }, + "cross-013": { + "answer": "Skiff runs Tern at the Performance preset, which saves 1.8 ms over Quality (2.1 ms vs 3.9 ms), per `dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3`.", + "tokens": 38711, + "fixed_tokens": 1106, + "processed_tokens": 338418, + "tool_calls": 13, + "turns": 13 + }, + "cross-014": { + "answer": 3, + "tokens": 5460, + "fixed_tokens": 1102, + "processed_tokens": 40290, + "tool_calls": 4, + "turns": 5 + }, + "cross-015": { + "answer": "Two Efforts are no longer active. Shader compile cache and build farm is completed (status only; I found no closing record saying every farm build reads from the cache). Co-op rollback netcode prototype was abandoned by dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b, which stops rollback because re-running tides on every rollback frame costs 4.8 ms on Skiff.", + "tokens": 46958, + "fixed_tokens": 1105, + "processed_tokens": 688261, + "tool_calls": 21, + "turns": 20 + }, + "cross-016": { + "answer": "16.7 ms", + "tokens": 48121, + "fixed_tokens": 1107, + "processed_tokens": 487831, + "tool_calls": 15, + "turns": 16 + }, + "cross-017": { + "answer": false, + "tokens": 15680, + "fixed_tokens": 1111, + "processed_tokens": 220459, + "tool_calls": 16, + "turns": 17 + }, + "netcode-001": { + "answer": "abandoned", + "tokens": 19769, + "fixed_tokens": 1106, + "processed_tokens": 134331, + "tool_calls": 8, + "turns": 9 + }, + "netcode-002": { + "answer": "netcode/rollback-archive", + "tokens": 17833, + "fixed_tokens": 1111, + "processed_tokens": 114698, + "tool_calls": 7, + "turns": 8 + }, + "netcode-003": { + "answer": "unknown", + "tokens": 43346, + "fixed_tokens": 1111, + "processed_tokens": 715784, + "tool_calls": 22, + "turns": 23 + }, + "netcode-004": { + "answer": "The team abandoned rollback netcode for co-op because re-running the tide simulation on every rolled-back frame costs 4.8 ms on Skiff, about a fifth of the 25 ms frame, and shrinking the window and quarter-resolution tides both failed (dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b). The reason depends on the world Decision dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e, which keeps tides deterministic.", + "tokens": 25179, + "fixed_tokens": 1111, + "processed_tokens": 325640, + "tool_calls": 15, + "turns": 16 + }, + "netcode-005": { + "answer": "unknown", + "tokens": 39699, + "fixed_tokens": 1112, + "processed_tokens": 486678, + "tool_calls": 16, + "turns": 17 + }, + "netcode-006": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 20165, + "fixed_tokens": 1109, + "processed_tokens": 144266, + "tool_calls": 8, + "turns": 9 + }, + "netcode-007": { + "answer": "dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b", + "tokens": 17878, + "fixed_tokens": 1111, + "processed_tokens": 114586, + "tool_calls": 7, + "turns": 8 + }, + "netcode-008": { + "answer": "dec-use-a-4-frame-rollback-window--pqe4hvbgmbp0dmh9", + "tokens": 66456, + "fixed_tokens": 1112, + "processed_tokens": 741200, + "tool_calls": 21, + "turns": 22 + }, + "netcode-009": { + "answer": "1.9 ms", + "tokens": 24679, + "fixed_tokens": 1106, + "processed_tokens": 300668, + "tool_calls": 14, + "turns": 15 + }, + "netcode-010": { + "answer": "dec-use-a-6-frame-rollback-window-with-1-frame-of-in--s53944zhkw103frk", + "tokens": 49976, + "fixed_tokens": 1110, + "processed_tokens": 680080, + "tool_calls": 18, + "turns": 19 + }, + "netcode-011": { + "answer": "con-co-op-peers-must-agree-on-tide-state-bit-for-bit--4v0sdbc3t4ywzr3g", + "tokens": 41024, + "fixed_tokens": 1108, + "processed_tokens": 709792, + "tool_calls": 22, + "turns": 23 + }, + "netcode-012": { + "answer": "unknown", + "tokens": 29923, + "fixed_tokens": 1113, + "processed_tokens": 357667, + "tool_calls": 15, + "turns": 16 + }, + "netcode-013": { + "answer": [ + "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "fnd-rollback-re-simulates-a-4-frame-window-in-1-1-ms--xa30a1cez2brphyf", + "rsk-re-simulation-cost-could-blow-the-skiff-frame-bu--sk2njsp170a3qwyf" + ], + "tokens": 30776, + "fixed_tokens": 1116, + "processed_tokens": 295083, + "tool_calls": 15, + "turns": 14 + }, + "netcode-014": { + "answer": [ + "iss-how-should-co-op-keep-up-to-four-players-in-sync--92t2zfsy3etfzr89", + "iss-late-join-during-a-tide-surge-desyncs-the-joinin--mya93va0n7s8g7cj" + ], + "tokens": 65083, + "fixed_tokens": 1107, + "processed_tokens": 905338, + "tool_calls": 24, + "turns": 25 + }, + "netcode-015": { + "answer": 2, + "tokens": 21878, + "fixed_tokens": 1108, + "processed_tokens": 152243, + "tool_calls": 9, + "turns": 10 + }, + "ports-001": { + "answer": "25 ms", + "tokens": 9722, + "fixed_tokens": 1108, + "processed_tokens": 115875, + "tool_calls": 10, + "turns": 11 + }, + "ports-002": { + "answer": "unknown", + "tokens": 69882, + "fixed_tokens": 1104, + "processed_tokens": 1255216, + "tool_calls": 26, + "turns": 25 + }, + "ports-003": { + "answer": "unknown", + "tokens": 23897, + "fixed_tokens": 1102, + "processed_tokens": 325684, + "tool_calls": 15, + "turns": 16 + }, + "ports-004": { + "answer": "accepted", + "tokens": 19753, + "fixed_tokens": 1112, + "processed_tokens": 126696, + "tool_calls": 8, + "turns": 9 + }, + "ports-005": { + "answer": "unknown", + "tokens": 61380, + "fixed_tokens": 1105, + "processed_tokens": 1517402, + "tool_calls": 35, + "turns": 36 + }, + "ports-034": { + "answer": "6 weeks", + "tokens": 36761, + "fixed_tokens": 1109, + "processed_tokens": 366997, + "tool_calls": 13, + "turns": 14 + }, + "ports-006": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 17232, + "fixed_tokens": 1103, + "processed_tokens": 112583, + "tool_calls": 7, + "turns": 8 + }, + "ports-007": { + "answer": "SKF-UI-031", + "tokens": 64594, + "fixed_tokens": 1106, + "processed_tokens": 2472667, + "tool_calls": 53, + "turns": 54 + }, + "ports-008": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 7713, + "fixed_tokens": 1107, + "processed_tokens": 72862, + "tool_calls": 7, + "turns": 8 + }, + "ports-009": { + "answer": "dec-load-skiff-fast-travel-without-a-loading-screen--v4qft5znysjv5f24", + "tokens": 61889, + "fixed_tokens": 1110, + "processed_tokens": 2782093, + "tool_calls": 77, + "turns": 78 + }, + "ports-010": { + "answer": "con-frame-budget--wt5rfqh5z444g2cg", + "tokens": 71923, + "fixed_tokens": 1106, + "processed_tokens": 2123584, + "tool_calls": 42, + "turns": 43 + }, + "ports-011": { + "answer": "99.6%", + "tokens": 67458, + "fixed_tokens": 1108, + "processed_tokens": 2541990, + "tool_calls": 56, + "turns": 56 + }, + "ports-012": { + "answer": "The ports team raised the Skiff dynamic resolution floor to 576p because saltglass caustics shimmered badly in motion at the 540p floor (visible in 7 of 10 dune views, versus 2 of 10 at 576p), and 576p still fit the 33.3 ms budget at 30 Hz (dec-raise-the-skiff-dynamic-resolution-floor-to-576p--fep9q7m08pp9f0bb). That decision was later superseded by dec-lower-the-skiff-dynamic-resolution-floor-to-504p--15esnt90gcn29enn, which lowered the floor to 504p behind Tern.", + "tokens": 42021, + "fixed_tokens": 1105, + "processed_tokens": 637299, + "tool_calls": 19, + "turns": 20 + }, + "ports-013": { + "answer": "Ship Tern as Lantern's only upscaler", + "tokens": 78033, + "fixed_tokens": 1105, + "processed_tokens": 1776819, + "tool_calls": 37, + "turns": 35 + }, + "ports-014": { + "answer": "dec-scale-skiff-resolution-dynamically-between-540p--jtytvgh2erd32x68", + "tokens": 21901, + "fixed_tokens": 1111, + "processed_tokens": 319466, + "tool_calls": 19, + "turns": 20 + }, + "ports-015": { + "answer": "textures", + "tokens": 24376, + "fixed_tokens": 1106, + "processed_tokens": 327226, + "tool_calls": 15, + "turns": 16 + }, + "ports-016": { + "answer": "dec-ship-prebuilt-pipeline-blobs-in-the-skiff-packag--3vjysjgy9qkvzebn", + "tokens": 24277, + "fixed_tokens": 1102, + "processed_tokens": 280608, + "tool_calls": 14, + "turns": 15 + }, + "ports-017": { + "answer": "unknown", + "tokens": 80205, + "fixed_tokens": 1112, + "processed_tokens": 3227118, + "tool_calls": 63, + "turns": 64 + }, + "ports-018": { + "answer": "504p", + "tokens": 23808, + "fixed_tokens": 1111, + "processed_tokens": 194668, + "tool_calls": 10, + "turns": 11 + }, + "ports-019": { + "answer": "unknown", + "tokens": 51121, + "fixed_tokens": 1104, + "processed_tokens": 1270926, + "tool_calls": 37, + "turns": 33 + }, + "ports-020": { + "answer": "fnd-skiff-sdk-1-2-takes-1-gb-back-from-games--a0ybdgjzw64hwfqa", + "tokens": 18872, + "fixed_tokens": 1102, + "processed_tokens": 159614, + "tool_calls": 9, + "turns": 10 + }, + "ports-021": { + "answer": "unknown", + "tokens": 57703, + "fixed_tokens": 1104, + "processed_tokens": 1124201, + "tool_calls": 29, + "turns": 30 + }, + "ports-022": { + "answer": "cit-skiff-sdk-1-2-release-notes--7gwwn8xxffrx6npe", + "tokens": 25328, + "fixed_tokens": 1111, + "processed_tokens": 287576, + "tool_calls": 13, + "turns": 14 + }, + "ports-023": { + "answer": "fnd-tern-costs-3-9-ms-at-quality-and-2-1-ms-at-perfo--dkbskbqfgfwtryq0", + "tokens": 26682, + "fixed_tokens": 1114, + "processed_tokens": 392576, + "tool_calls": 17, + "turns": 18 + }, + "ports-024": { + "answer": "fnd-skiff-firmware-1-2-rejects-pipeline-blobs-built--pernvgb2he7m3z3y", + "tokens": 45835, + "fixed_tokens": 1103, + "processed_tokens": 885066, + "tool_calls": 25, + "turns": 26 + }, + "ports-025": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 30400, + "fixed_tokens": 1101, + "processed_tokens": 403303, + "tool_calls": 16, + "turns": 17 + }, + "ports-026": { + "answer": "912 MB", + "tokens": 51753, + "fixed_tokens": 1104, + "processed_tokens": 993463, + "tool_calls": 25, + "turns": 26 + }, + "ports-027": { + "answer": [ + "dec-stream-skiff-world-tiles-in-2-mb-reads--3jz9w9gvhhc7eppw", + "rsk-tide-streaming-outruns-skiff-storage--6q8s3b383z2tyeyb" + ], + "tokens": 61038, + "fixed_tokens": 1098, + "processed_tokens": 1560641, + "tool_calls": 39, + "turns": 38 + }, + "ports-028": { + "answer": 2, + "tokens": 7223, + "fixed_tokens": 1101, + "processed_tokens": 51818, + "tool_calls": 5, + "turns": 6 + }, + "ports-029": { + "answer": null, + "tokens": 176375, + "fixed_tokens": 1109, + "processed_tokens": 3926982, + "tool_calls": 77, + "turns": 45 + }, + "ports-030": { + "answer": [ + "dec-cap-tide-storm-particles-at-6-000-on-skiff--resze52afjxyxys1", + "dec-drop-skiff-to-30-hz-while-a-tide-storm-is-on-scr--72vx1c8s4jrs6axc" + ], + "tokens": 7397, + "fixed_tokens": 1105, + "processed_tokens": 62163, + "tool_calls": 6, + "turns": 7 + }, + "ports-031": { + "answer": [ + "iss-skiff-cert-docked-30-hz-judder-on-some-tvs--9nq2xg9mzash474g", + "iss-meridian-cert-activity-cards-missing-for-two-bio--k214cmfra7pyvvzq" + ], + "tokens": 37045, + "fixed_tokens": 1099, + "processed_tokens": 626670, + "tool_calls": 23, + "turns": 24 + }, + "ports-032": { + "answer": 2, + "tokens": 6673, + "fixed_tokens": 1106, + "processed_tokens": 61540, + "tool_calls": 6, + "turns": 7 + }, + "ports-033": { + "answer": "unknown", + "tokens": 20177, + "fixed_tokens": 1107, + "processed_tokens": 118720, + "tool_calls": 7, + "turns": 8 + }, + "render-001": { + "answer": "7.4 ms", + "tokens": 24124, + "fixed_tokens": 1106, + "processed_tokens": 171353, + "tool_calls": 9, + "turns": 10 + }, + "render-002": { + "answer": "3.4x", + "tokens": 26855, + "fixed_tokens": 1100, + "processed_tokens": 339945, + "tool_calls": 15, + "turns": 16 + }, + "render-003": { + "answer": "Tomasz Brandt", + "tokens": 32737, + "fixed_tokens": 1099, + "processed_tokens": 214189, + "tool_calls": 10, + "turns": 11 + }, + "render-004": { + "answer": "rejected", + "tokens": 34314, + "fixed_tokens": 1104, + "processed_tokens": 490839, + "tool_calls": 18, + "turns": 19 + }, + "render-005": { + "answer": true, + "tokens": 25659, + "fixed_tokens": 1100, + "processed_tokens": 253011, + "tool_calls": 12, + "turns": 13 + }, + "render-006": { + "answer": "unknown", + "tokens": 30646, + "fixed_tokens": 1102, + "processed_tokens": 575561, + "tool_calls": 22, + "turns": 23 + }, + "render-007": { + "answer": "2026-07-24", + "tokens": 46020, + "fixed_tokens": 1106, + "processed_tokens": 379008, + "tool_calls": 12, + "turns": 13 + }, + "render-008": { + "answer": "2", + "tokens": 29427, + "fixed_tokens": 1098, + "processed_tokens": 281859, + "tool_calls": 12, + "turns": 13 + }, + "render-009": { + "answer": "32 steps", + "tokens": 21116, + "fixed_tokens": 1108, + "processed_tokens": 190816, + "tool_calls": 10, + "turns": 11 + }, + "render-010": { + "answer": "unknown", + "tokens": 22025, + "fixed_tokens": 1100, + "processed_tokens": 154744, + "tool_calls": 9, + "turns": 10 + }, + "render-011": { + "answer": "fnd-sdf-gi-costs-2-4-ms-on-meridian--2w6c2pc5zmcsyjre", + "tokens": 37751, + "fixed_tokens": 1108, + "processed_tokens": 579181, + "tool_calls": 30, + "turns": 26 + }, + "render-012": { + "answer": "deferred", + "tokens": 32797, + "fixed_tokens": 1096, + "processed_tokens": 625120, + "tool_calls": 25, + "turns": 25 + }, + "render-013": { + "answer": "16-bit float per channel", + "tokens": 59872, + "fixed_tokens": 1107, + "processed_tokens": 1097003, + "tool_calls": 25, + "turns": 26 + }, + "render-014": { + "answer": "52%", + "tokens": 30424, + "fixed_tokens": 1103, + "processed_tokens": 472072, + "tool_calls": 21, + "turns": 22 + }, + "render-015": { + "answer": "unknown", + "tokens": 40322, + "fixed_tokens": 1103, + "processed_tokens": 1221463, + "tool_calls": 47, + "turns": 48 + }, + "render-016": { + "answer": "1.2 km", + "tokens": 22013, + "fixed_tokens": 1106, + "processed_tokens": 154953, + "tool_calls": 9, + "turns": 10 + }, + "render-017": { + "answer": "unknown", + "tokens": 45258, + "fixed_tokens": 1101, + "processed_tokens": 895169, + "tool_calls": 25, + "turns": 26 + }, + "render-018": { + "answer": "unknown", + "tokens": 50283, + "fixed_tokens": 1102, + "processed_tokens": 1461539, + "tool_calls": 42, + "turns": 42 + }, + "render-019": { + "answer": "dec-use-the-custom-brine-curve-for-tonemapping--p735g6wf2njare86", + "tokens": 30167, + "fixed_tokens": 1107, + "processed_tokens": 224943, + "tool_calls": 11, + "turns": 11 + }, + "render-020": { + "answer": "fnd-screen-space-refraction-loses-off-screen-content--k8zvmmdee7wg2fjy", + "tokens": 27168, + "fixed_tokens": 1103, + "processed_tokens": 637421, + "tool_calls": 31, + "turns": 32 + }, + "render-021": { + "answer": "dec-adopt-a-visibility-buffer-for-all-opaque-geometr--68vchzqgd1avw8q4", + "tokens": 8459, + "fixed_tokens": 1099, + "processed_tokens": 83980, + "tool_calls": 8, + "turns": 9 + }, + "render-022": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 8767, + "fixed_tokens": 1100, + "processed_tokens": 86613, + "tool_calls": 8, + "turns": 9 + }, + "render-023": { + "answer": "dec-capture-the-cliffs-and-dunes-benchmarks-on-every--3pqyr66yh4v6q676", + "tokens": 63944, + "fixed_tokens": 1104, + "processed_tokens": 738659, + "tool_calls": 19, + "turns": 18 + }, + "render-024": { + "answer": "unknown", + "tokens": 27249, + "fixed_tokens": 1100, + "processed_tokens": 297269, + "tool_calls": 15, + "turns": 14 + }, + "render-025": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-after-fp16-pa--p4gcmzw2a4bvzjev", + "tokens": 49589, + "fixed_tokens": 1106, + "processed_tokens": 1836418, + "tool_calls": 63, + "turns": 64 + }, + "render-026": { + "answer": "dec-ship-four-tern-quality-presets--vjw18d82dx5gath8", + "tokens": 27872, + "fixed_tokens": 1101, + "processed_tokens": 310824, + "tool_calls": 14, + "turns": 15 + }, + "render-027": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 11350, + "fixed_tokens": 1100, + "processed_tokens": 78101, + "tool_calls": 7, + "turns": 8 + }, + "render-028": { + "answer": "dec-move-to-gpu-driven-two-phase-occlusion-culling--6dre481mabxsw6ee", + "tokens": 26200, + "fixed_tokens": 1109, + "processed_tokens": 325351, + "tool_calls": 15, + "turns": 16 + }, + "render-029": { + "answer": "fnd-vsm-page-pool-overflows-during-spring-tides--m2p8e89hhdmh0z8a", + "tokens": 12957, + "fixed_tokens": 1104, + "processed_tokens": 106718, + "tool_calls": 8, + "turns": 9 + }, + "render-030": { + "answer": "dec-use-screen-space-refraction-for-saltglass--vknzksgdrgr7zkg6", + "tokens": 37860, + "fixed_tokens": 1106, + "processed_tokens": 678769, + "tool_calls": 24, + "turns": 25 + }, + "render-031": { + "answer": "dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9: alpha-tested foliage cost 2.1 ms in the visibility buffer versus 0.8 ms in forward+, and saltglass refracts the already-resolved scene so it cannot live in the buffer.", + "tokens": 21235, + "fixed_tokens": 1100, + "processed_tokens": 213334, + "tool_calls": 11, + "turns": 12 + }, + "render-032": { + "answer": "The team chose Tern because Halcyon has no Skiff build, and shipping Halcyon on two platforms would break the one-upscaler constraint; owning the Tern code also suits saltglass refraction and caustics better than vendor tickets and a license negotiation would (dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e).", + "tokens": 31531, + "fixed_tokens": 1103, + "processed_tokens": 313447, + "tool_calls": 15, + "turns": 14 + }, + "render-033": { + "answer": "128 MB", + "tokens": 26157, + "fixed_tokens": 1108, + "processed_tokens": 206431, + "tool_calls": 10, + "turns": 11 + }, + "render-034": { + "answer": "PC workstation (Meridian clock profile)", + "tokens": 26485, + "fixed_tokens": 1098, + "processed_tokens": 591390, + "tool_calls": 30, + "turns": 31 + }, + "render-035": { + "answer": "5912 pages", + "tokens": 28159, + "fixed_tokens": 1100, + "processed_tokens": 328778, + "tool_calls": 14, + "turns": 15 + }, + "render-036": { + "answer": "unknown", + "tokens": 36184, + "fixed_tokens": 1105, + "processed_tokens": 936842, + "tool_calls": 32, + "turns": 33 + }, + "render-037": { + "answer": "unknown", + "tokens": 30799, + "fixed_tokens": 1100, + "processed_tokens": 564994, + "tool_calls": 26, + "turns": 27 + }, + "render-038": { + "answer": "dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9", + "tokens": 27788, + "fixed_tokens": 1107, + "processed_tokens": 347539, + "tool_calls": 15, + "turns": 16 + }, + "render-039": { + "answer": "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "tokens": 33230, + "fixed_tokens": 1107, + "processed_tokens": 288955, + "tool_calls": 12, + "turns": 13 + }, + "render-040": { + "answer": "fnd-lantern-at-alpha--ckpd9ecvdjqbcr7c", + "tokens": 20862, + "fixed_tokens": 1102, + "processed_tokens": 138289, + "tool_calls": 8, + "turns": 9 + }, + "render-041": { + "answer": "4.7 ms", + "tokens": 36867, + "fixed_tokens": 1101, + "processed_tokens": 845420, + "tool_calls": 30, + "turns": 31 + }, + "render-042": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 16789, + "fixed_tokens": 1104, + "processed_tokens": 167774, + "tool_calls": 11, + "turns": 12 + }, + "render-043": { + "answer": "fnd-meshlet-cone-culling-rejects-9-percent-of-triang--t6e5x7trzedex105", + "tokens": 49330, + "fixed_tokens": 1102, + "processed_tokens": 1438431, + "tool_calls": 42, + "turns": 43 + }, + "render-044": { + "answer": "dec-screen-space-refraction-with-a-ray-marched-fallb--2639xtpyrfrvbx1r", + "tokens": 14357, + "fixed_tokens": 1102, + "processed_tokens": 158974, + "tool_calls": 12, + "turns": 13 + }, + "render-045": { + "answer": "false, dec-brine-curve-with-an-hdr-shoulder-at-800-nits--fpznx0cnj4avsxk1", + "tokens": 50225, + "fixed_tokens": 1105, + "processed_tokens": 628000, + "tool_calls": 19, + "turns": 20 + }, + "render-046": { + "answer": "The 3.1 ms cost finding that rejected them (fnd-virtual-shadow-maps-cost-3-1-ms-on-meridian--r6x0sdfecms7j1yg) was a capture error: the debug flag discarded cached pages. With caching, VSM costs 1.4 ms, less than four cascades' 1.6 ms, and CSM's cascade seams through saltglass remain unsolved; see dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4.", + "tokens": 15262, + "fixed_tokens": 1103, + "processed_tokens": 221677, + "tool_calls": 16, + "turns": 17 + }, + "render-047": { + "answer": "Because the release-build capture puts the visibility buffer at 7.4 ms on Skiff against 8.1 ms for forward+ (6.3 ms with material resolve at half rate), so the forward+ fallback is no longer worth keeping, and keeping two opaque paths cost about one engineer-week a month in fixes. dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 19635, + "fixed_tokens": 1104, + "processed_tokens": 264597, + "tool_calls": 16, + "turns": 17 + }, + "render-048": { + "answer": "unknown", + "tokens": 45223, + "fixed_tokens": 1106, + "processed_tokens": 1109961, + "tool_calls": 34, + "turns": 34 + }, + "render-049": { + "answer": "dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4", + "tokens": 28761, + "fixed_tokens": 1094, + "processed_tokens": 586471, + "tool_calls": 30, + "turns": 31 + }, + "render-050": { + "answer": "con-frame-budget--fzdrbxns6dmrrmy3", + "tokens": 9915, + "fixed_tokens": 1097, + "processed_tokens": 57223, + "tool_calls": 5, + "turns": 6 + }, + "render-051": { + "answer": "iss-tern-ghosting-on-moving-saltglass--2yr3q6zf476wajqp", + "tokens": 30966, + "fixed_tokens": 1100, + "processed_tokens": 302819, + "tool_calls": 13, + "turns": 14 + }, + "render-052": { + "answer": "61 percent", + "tokens": 34795, + "fixed_tokens": 1099, + "processed_tokens": 547748, + "tool_calls": 23, + "turns": 22 + }, + "render-053": { + "answer": "every frame", + "tokens": 28380, + "fixed_tokens": 1098, + "processed_tokens": 441910, + "tool_calls": 18, + "turns": 19 + }, + "render-054": { + "answer": "debug-gpu-validation", + "tokens": 32192, + "fixed_tokens": 1101, + "processed_tokens": 557692, + "tool_calls": 21, + "turns": 22 + }, + "render-055": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-with-the-reac--81tcpzvhwzcy6wyn", + "tokens": 48364, + "fixed_tokens": 1101, + "processed_tokens": 624606, + "tool_calls": 20, + "turns": 20 + }, + "render-056": { + "answer": "208,000 meshlets", + "tokens": 29649, + "fixed_tokens": 1106, + "processed_tokens": 531955, + "tool_calls": 21, + "turns": 22 + }, + "render-057": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 33204, + "fixed_tokens": 1101, + "processed_tokens": 516375, + "tool_calls": 20, + "turns": 21 + }, + "render-058": { + "answer": "iss-which-upscaler-should-lantern-ship--g8c2y9tsr856tn2v", + "tokens": 57570, + "fixed_tokens": 1101, + "processed_tokens": 1703407, + "tool_calls": 44, + "turns": 45 + }, + "render-059": { + "answer": "fnd-sdf-gi-costs-4-7-ms-on-meridian--sm19rt2f274ajsvr", + "tokens": 35419, + "fixed_tokens": 1098, + "processed_tokens": 460874, + "tool_calls": 17, + "turns": 18 + }, + "render-060": { + "answer": "0.31", + "tokens": 34322, + "fixed_tokens": 1100, + "processed_tokens": 632382, + "tool_calls": 24, + "turns": 25 + }, + "render-061": { + "answer": "con-ui-composites-at-a-fixed-203-nits-in-hdr--de6aft7cvhdezn5s", + "tokens": 42002, + "fixed_tokens": 1102, + "processed_tokens": 658478, + "tool_calls": 23, + "turns": 21 + }, + "render-062": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 23787, + "fixed_tokens": 1102, + "processed_tokens": 148015, + "tool_calls": 8, + "turns": 9 + }, + "render-063": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 13839, + "fixed_tokens": 1110, + "processed_tokens": 104590, + "tool_calls": 10, + "turns": 9 + }, + "render-064": { + "answer": "dec-use-a-cached-shadow-atlas-for-local-lights--8p6f2v774he38t5j", + "tokens": 30974, + "fixed_tokens": 1109, + "processed_tokens": 333723, + "tool_calls": 16, + "turns": 15 + }, + "render-065": { + "answer": "dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e", + "tokens": 69673, + "fixed_tokens": 1102, + "processed_tokens": 1483644, + "tool_calls": 30, + "turns": 31 + }, + "render-066": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 26917, + "fixed_tokens": 1105, + "processed_tokens": 407623, + "tool_calls": 18, + "turns": 19 + }, + "render-067": { + "answer": 4, + "tokens": 22032, + "fixed_tokens": 1101, + "processed_tokens": 151536, + "tool_calls": 9, + "turns": 10 + }, + "render-068": { + "answer": [ + "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "dec-lock-caustic-photon-jitter-to-the-tide-phase--85x10k4h11t8nx7c", + "dec-use-the-probe-grid-alone-in-120-hz-performance-m--6dzm1nt40pjfpyae", + "dec-trace-sdf-gi-at-quarter-resolution-in-performanc--06cqt3mqn3fg5h6e", + "dec-read-paper-white-from-the-meridian-system-settin--7crbx73gqw1hn7xq", + "dec-add-an-in-game-paper-white-slider-from-80-to-400--1phfvxby983z8tjf" + ], + "tokens": 23436, + "fixed_tokens": 1099, + "processed_tokens": 124760, + "tool_calls": 7, + "turns": 8 + }, + "render-069": { + "answer": "iss-shadow-pages-run-out-at-spring-low-tide--f4fgy6b44jbkrsmh", + "tokens": 26163, + "fixed_tokens": 1108, + "processed_tokens": 189932, + "tool_calls": 10, + "turns": 11 + }, + "render-070": { + "answer": null, + "tokens": 115829, + "fixed_tokens": 1107, + "processed_tokens": 4071252, + "tool_calls": 85, + "turns": 62 + }, + "render-071": { + "answer": 3, + "tokens": 12017, + "fixed_tokens": 1099, + "processed_tokens": 98652, + "tool_calls": 10, + "turns": 9 + }, + "render-072": { + "answer": [ + "rsk-sdf-volumes-outgrow-the-render-target-memory-bud--xfd2pfztvskdcd2c", + "rsk-artists-cannot-match-concept-paintings-under-the--6cfh9vkamn0kbtcx" + ], + "tokens": 9517, + "fixed_tokens": 1107, + "processed_tokens": 56128, + "tool_calls": 5, + "turns": 6 + }, + "render-073": { + "answer": [ + "rsk-vsm-page-pool-overflows-when-tides-expose-the-fl--x1r161nrdngra4ty", + "rsk-gi-cannot-fit-the-120-hz-performance-mode--tqt3fj839dqfq5s7", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 9515, + "fixed_tokens": 1106, + "processed_tokens": 56164, + "tool_calls": 5, + "turns": 6 + }, + "render-074": { + "answer": 5, + "tokens": 8235, + "fixed_tokens": 1107, + "processed_tokens": 66367, + "tool_calls": 6, + "turns": 7 + }, + "render-075": { + "answer": 3, + "tokens": 33578, + "fixed_tokens": 1103, + "processed_tokens": 658364, + "tool_calls": 24, + "turns": 25 + }, + "render-076": { + "answer": true, + "tokens": 16520, + "fixed_tokens": 1108, + "processed_tokens": 224016, + "tool_calls": 16, + "turns": 17 + }, + "render-077": { + "answer": true, + "tokens": 34296, + "fixed_tokens": 1099, + "processed_tokens": 307832, + "tool_calls": 23, + "turns": 16 + }, + "render-078": { + "answer": [ + "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 36841, + "fixed_tokens": 1110, + "processed_tokens": 454623, + "tool_calls": 16, + "turns": 17 + }, + "render-079": { + "answer": 5, + "tokens": 51576, + "fixed_tokens": 1103, + "processed_tokens": 805010, + "tool_calls": 25, + "turns": 25 + }, + "render-080": { + "answer": "6", + "tokens": 29923, + "fixed_tokens": 1105, + "processed_tokens": 358523, + "tool_calls": 15, + "turns": 16 + }, + "render-081": { + "answer": null, + "tokens": 66534, + "fixed_tokens": 1100, + "processed_tokens": 3114233, + "tool_calls": 77, + "turns": 77 + }, + "render-082": { + "answer": "false", + "tokens": 67541, + "fixed_tokens": 1107, + "processed_tokens": 2833002, + "tool_calls": 64, + "turns": 65 + }, + "render-083": { + "answer": null, + "tokens": 96132, + "fixed_tokens": 1099, + "processed_tokens": 3654902, + "tool_calls": 64, + "turns": 64 + }, + "tools-001": { + "answer": "completed", + "tokens": 19151, + "fixed_tokens": 1103, + "processed_tokens": 111647, + "tool_calls": 7, + "turns": 8 + }, + "tools-002": { + "answer": "47 minutes", + "tokens": 18124, + "fixed_tokens": 1103, + "processed_tokens": 116071, + "tool_calls": 7, + "turns": 8 + }, + "tools-003": { + "answer": "6 min 40 s", + "tokens": 39838, + "fixed_tokens": 1112, + "processed_tokens": 992968, + "tool_calls": 32, + "turns": 33 + }, + "tools-004": { + "answer": "unknown", + "tokens": 11370, + "fixed_tokens": 1105, + "processed_tokens": 126857, + "tool_calls": 10, + "turns": 11 + }, + "tools-005": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 33304, + "fixed_tokens": 1113, + "processed_tokens": 226452, + "tool_calls": 10, + "turns": 11 + }, + "tools-006": { + "answer": "dec-build-a-content-addressed-shader-compile-cache-s--e65rq9439k7xz5zh", + "tokens": 19267, + "fixed_tokens": 1105, + "processed_tokens": 120422, + "tool_calls": 7, + "turns": 8 + }, + "tools-007": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 61875, + "fixed_tokens": 1110, + "processed_tokens": 1333936, + "tool_calls": 28, + "turns": 29 + }, + "tools-008": { + "answer": "dec-key-the-shader-cache-on-raw-source-text-and-comp--hp6awmx3gq5rn9nc", + "tokens": 29284, + "fixed_tokens": 1103, + "processed_tokens": 286408, + "tool_calls": 12, + "turns": 13 + }, + "tools-009": { + "answer": "92.4%", + "tokens": 22402, + "fixed_tokens": 1111, + "processed_tokens": 234991, + "tool_calls": 12, + "turns": 13 + }, + "tools-010": { + "answer": "dec-cap-each-farm-agent-at-6-parallel-shader-compile--q5w4sfjm73tsj8x8", + "tokens": 30513, + "fixed_tokens": 1103, + "processed_tokens": 530444, + "tool_calls": 21, + "turns": 22 + }, + "tools-011": { + "answer": "fnd-two-compiler-builds-with-the-same-version-string--gmgr4ha0ndxqmcn9", + "tokens": 39745, + "fixed_tokens": 1107, + "processed_tokens": 691560, + "tool_calls": 32, + "turns": 24 + }, + "tools-012": { + "answer": 3, + "tokens": 20561, + "fixed_tokens": 1102, + "processed_tokens": 113926, + "tool_calls": 7, + "turns": 8 + }, + "tools-013": { + "answer": 6, + "tokens": 93668, + "fixed_tokens": 1102, + "processed_tokens": 976286, + "tool_calls": 20, + "turns": 20 + }, + "tools-014": { + "answer": false, + "tokens": 31350, + "fixed_tokens": 1106, + "processed_tokens": 277634, + "tool_calls": 14, + "turns": 13 + }, + "tools-015": { + "answer": [ + "iss-whether-the-artist-sandbox-should-be-split-into--ek47ca2vgmrjanp8", + "iss-whether-the-artist-sandbox-should-load-on-a-work--73j819anqct8etfn", + "iss-whether-the-artist-sandbox-should-skip-work-when--11tqzws72gms2h8x", + "iss-whether-the-asset-search-should-load-on-a-worker--80b27kqh36rpafdt", + "iss-whether-the-asset-search-should-load-on-a-worker--f9gpv3z522x6p6h4", + "iss-whether-the-asset-search-should-ship-in-the-firs--hsb2avhx613prxew", + "iss-whether-the-audio-bank-packer-needs-its-own-memo--qjrqrseast8d5mhf", + "iss-whether-the-audio-bank-packer-should-load-on-a-w--sqzp2rwp6s01w3b0", + "iss-whether-the-changelog-bot-needs-its-own-memory-p--mw8zejz6hk5m60x0", + "iss-whether-the-changelog-bot-should-load-on-a-worke--jgtdn2f02p8fx5nr", + "iss-whether-the-changelog-bot-should-skip-work-when--p2vrp5d682p3jrc3", + "iss-whether-the-crash-symbol-server-should-load-on-a--qvyphhgnsnm2cvqb", + "iss-whether-the-crash-symbol-server-should-ship-in-t--bhq91ajavah18wsk", + "iss-whether-the-doc-site-should-load-on-a-worker-thr--qxswe0pg52ekky0z", + "iss-whether-the-doc-site-should-ship-in-the-first-pa--wcdkq5zab7dxe5yj", + "iss-whether-the-level-editor-autosave-needs-its-own--rkvmnbqe7w644zkr", + "iss-whether-the-level-editor-autosave-should-be-cach--radcjs4hyvh7p55q", + "iss-whether-the-level-editor-autosave-should-be-spli--ncnwh0csw671fs0q", + "iss-whether-the-level-editor-autosave-should-skip-wo--fyysrcv6fvbsvcsy", + "iss-whether-the-license-scanner-should-be-cached-bet--nqhd89czbfat9rya", + "iss-whether-the-patch-differ-should-be-cached-betwee--28b5zqb1ketwzf0r", + "iss-whether-the-patch-differ-should-load-on-a-worker--cysv4n56q1qn7t2k", + "iss-whether-the-review-bot-needs-its-own-memory-pool--pg3m8g9gn3y6v3kq", + "iss-whether-the-review-bot-should-load-on-a-worker-t--rtsjwvsksjdq15v1", + "iss-whether-the-test-runner-needs-its-own-memory-poo--bs6arq0qv37d6wmx", + "iss-whether-the-test-runner-should-load-on-a-worker--aygay1r0qj4mzf78", + "iss-whether-the-texture-viewer-should-be-split-into--1jhgyqb70h5x7m7b", + "rsk-shader-cache-storage-could-outgrow-the-4-tb-farm--76y9649mwrx7aavd" + ], + "tokens": 78062, + "fixed_tokens": 1111, + "processed_tokens": 918222, + "tool_calls": 21, + "turns": 21 + }, + "tools-016": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 38279, + "fixed_tokens": 1103, + "processed_tokens": 934146, + "tool_calls": 31, + "turns": 32 + }, + "world-001": { + "answer": "64 KB", + "tokens": 25486, + "fixed_tokens": 1098, + "processed_tokens": 256280, + "tool_calls": 12, + "turns": 13 + }, + "world-002": { + "answer": "5.1 ms", + "tokens": 27673, + "fixed_tokens": 1107, + "processed_tokens": 417193, + "tool_calls": 18, + "turns": 19 + }, + "world-003": { + "answer": "one crash per 900 script waits", + "tokens": 46937, + "fixed_tokens": 1104, + "processed_tokens": 569746, + "tool_calls": 17, + "turns": 18 + }, + "world-004": { + "answer": "4100 file opens", + "tokens": 38223, + "fixed_tokens": 1100, + "processed_tokens": 869882, + "tool_calls": 29, + "turns": 30 + }, + "world-005": { + "answer": "11 m", + "tokens": 40987, + "fixed_tokens": 1098, + "processed_tokens": 687444, + "tool_calls": 22, + "turns": 23 + }, + "world-006": { + "answer": "unknown", + "tokens": 41083, + "fixed_tokens": 1100, + "processed_tokens": 796958, + "tool_calls": 26, + "turns": 27 + }, + "world-007": { + "answer": "260 MB", + "tokens": 53546, + "fixed_tokens": 1100, + "processed_tokens": 1276106, + "tool_calls": 32, + "turns": 33 + }, + "world-008": { + "answer": "10 Hz", + "tokens": 81863, + "fixed_tokens": 1102, + "processed_tokens": 1624191, + "tool_calls": 30, + "turns": 31 + }, + "world-009": { + "answer": "unknown", + "tokens": 81245, + "fixed_tokens": 1108, + "processed_tokens": 1477680, + "tool_calls": 28, + "turns": 29 + }, + "world-010": { + "answer": "29 GB/s", + "tokens": 28260, + "fixed_tokens": 1100, + "processed_tokens": 258353, + "tool_calls": 12, + "turns": 13 + }, + "world-011": { + "answer": "Closed as wontfix: the Name component was not shipped in release builds because it would cost 57 MB at full population. Instead, the crash pipeline maps entity numbers back to names from the build symbol file. Resolving record: fnd-shipping-entity-names-costs-57-mb-at-full-popula--6w7gjvy7wdmdx9qx", + "tokens": 25166, + "fixed_tokens": 1103, + "processed_tokens": 525894, + "tool_calls": 28, + "turns": 29 + }, + "world-012": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 22042, + "fixed_tokens": 1101, + "processed_tokens": 141134, + "tool_calls": 8, + "turns": 9 + }, + "world-013": { + "answer": "unknown", + "tokens": 31361, + "fixed_tokens": 1103, + "processed_tokens": 762445, + "tool_calls": 32, + "turns": 33 + }, + "world-014": { + "answer": "unknown", + "tokens": 46538, + "fixed_tokens": 1102, + "processed_tokens": 1370630, + "tool_calls": 38, + "turns": 39 + }, + "world-015": { + "answer": "3.0 ms", + "tokens": 85457, + "fixed_tokens": 1103, + "processed_tokens": 463478, + "tool_calls": 21, + "turns": 12 + }, + "world-016": { + "answer": "unknown", + "tokens": 29681, + "fixed_tokens": 1104, + "processed_tokens": 498183, + "tool_calls": 20, + "turns": 20 + }, + "world-017": { + "answer": "dec-partition-the-world-into-128-m-cells--372b6czbjf29ymaw", + "tokens": 36081, + "fixed_tokens": 1105, + "processed_tokens": 739708, + "tool_calls": 26, + "turns": 27 + }, + "world-018": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 19561, + "fixed_tokens": 1110, + "processed_tokens": 129993, + "tool_calls": 8, + "turns": 9 + }, + "world-019": { + "answer": "fnd-tide-priority-reads-starve-player-proximity-load--by0d7b0bk39xz3rk", + "tokens": 45665, + "fixed_tokens": 1102, + "processed_tokens": 785189, + "tool_calls": 23, + "turns": 24 + }, + "world-020": { + "answer": "dec-schedule-streaming-reads-by-deadline-instead-of--q2x5aknabgcssdj2", + "tokens": 9158, + "fixed_tokens": 1106, + "processed_tokens": 67912, + "tool_calls": 6, + "turns": 7 + }, + "world-021": { + "answer": "dec-cut-navigation-tiles-to-64-m-so-three-fit-along--snvepk8ygeb3qsjs", + "tokens": 13324, + "fixed_tokens": 1103, + "processed_tokens": 108745, + "tool_calls": 8, + "turns": 9 + }, + "world-022": { + "answer": "fnd-41-of-alpha-crash-reports-have-no-usable-stack--35a50b07ferhf13m", + "tokens": 29949, + "fixed_tokens": 1103, + "processed_tokens": 295302, + "tool_calls": 13, + "turns": 14 + }, + "world-023": { + "answer": "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "tokens": 42244, + "fixed_tokens": 1101, + "processed_tokens": 642727, + "tool_calls": 21, + "turns": 22 + }, + "world-024": { + "answer": "dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew", + "tokens": 30580, + "fixed_tokens": 1102, + "processed_tokens": 313029, + "tool_calls": 13, + "turns": 14 + }, + "world-025": { + "answer": "con-streaming-pool-budget--kca88sgh2j6sxvd5", + "tokens": 11048, + "fixed_tokens": 1100, + "processed_tokens": 69276, + "tool_calls": 6, + "turns": 7 + }, + "world-026": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 55438, + "fixed_tokens": 1109, + "processed_tokens": 1742758, + "tool_calls": 51, + "turns": 52 + }, + "world-027": { + "answer": "160 MB/s", + "tokens": 43764, + "fixed_tokens": 1105, + "processed_tokens": 431777, + "tool_calls": 14, + "turns": 15 + }, + "world-028": { + "answer": "unknown", + "tokens": 35842, + "fixed_tokens": 1107, + "processed_tokens": 846634, + "tool_calls": 34, + "turns": 35 + }, + "world-029": { + "answer": "dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h", + "tokens": 25240, + "fixed_tokens": 1102, + "processed_tokens": 237225, + "tool_calls": 12, + "turns": 13 + }, + "world-030": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 34131, + "fixed_tokens": 1099, + "processed_tokens": 1038908, + "tool_calls": 44, + "turns": 45 + }, + "world-031": { + "answer": "1.5 GB/s", + "tokens": 54898, + "fixed_tokens": 1101, + "processed_tokens": 2166881, + "tool_calls": 64, + "turns": 65 + }, + "world-032": { + "answer": "fnd-float-tide-heights-drift-apart-between-pc-and-me--7jb0fg4b5xzr7q5e", + "tokens": 45275, + "fixed_tokens": 1102, + "processed_tokens": 1186220, + "tool_calls": 35, + "turns": 36 + }, + "world-033": { + "answer": "dec-save-player-changes-as-deltas-against-the-tide-f--n38scy1yaarsr8cw", + "tokens": 36934, + "fixed_tokens": 1106, + "processed_tokens": 869536, + "tool_calls": 36, + "turns": 37 + }, + "world-034": { + "answer": "unknown", + "tokens": 38561, + "fixed_tokens": 1105, + "processed_tokens": 851043, + "tool_calls": 31, + "turns": 32 + }, + "world-035": { + "answer": false, + "tokens": 30764, + "fixed_tokens": 1108, + "processed_tokens": 336531, + "tool_calls": 14, + "turns": 15 + }, + "world-036": { + "answer": "Kestrel retired fibers because 41% of alpha crash reports had no usable stack, which blocked crash triage for beta, and a thread pool with continuation waits kept workers 93% busy during tide turns, close to the 95% fibers gave (dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h).", + "tokens": 25325, + "fixed_tokens": 1102, + "processed_tokens": 260217, + "tool_calls": 12, + "turns": 13 + }, + "world-037": { + "answer": "fnd-hot-reload-drops-suspended-gull-coroutines--va0fn0svryzwbzws", + "tokens": 25611, + "fixed_tokens": 1104, + "processed_tokens": 418322, + "tool_calls": 22, + "turns": 23 + }, + "world-038": { + "answer": "912 archetypes", + "tokens": 24510, + "fixed_tokens": 1102, + "processed_tokens": 275115, + "tool_calls": 13, + "turns": 14 + }, + "world-039": { + "answer": "1.2 GB/s", + "tokens": 29993, + "fixed_tokens": 1109, + "processed_tokens": 551880, + "tool_calls": 22, + "turns": 23 + }, + "world-040": { + "answer": "unknown", + "tokens": 31022, + "fixed_tokens": 1101, + "processed_tokens": 596908, + "tool_calls": 24, + "turns": 25 + }, + "world-041": { + "answer": "61%", + "tokens": 34848, + "fixed_tokens": 1106, + "processed_tokens": 342381, + "tool_calls": 13, + "turns": 14 + }, + "world-042": { + "answer": "5.4 s", + "tokens": 32761, + "fixed_tokens": 1096, + "processed_tokens": 422571, + "tool_calls": 16, + "turns": 17 + }, + "world-043": { + "answer": "5.8 minutes", + "tokens": 31268, + "fixed_tokens": 1108, + "processed_tokens": 351651, + "tool_calls": 14, + "turns": 15 + }, + "world-044": { + "answer": "dec-move-engine-jobs-to-fibers--qjkgnm8cex8ydzfq", + "tokens": 27246, + "fixed_tokens": 1111, + "processed_tokens": 316069, + "tool_calls": 14, + "turns": 15 + }, + "world-045": { + "answer": "dec-run-engine-jobs-on-fibers-and-pin-gull-scripts-t--rw28cbqfd07rd9te", + "tokens": 27450, + "fixed_tokens": 1104, + "processed_tokens": 291979, + "tool_calls": 13, + "turns": 14 + }, + "world-046": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 11741, + "fixed_tokens": 1109, + "processed_tokens": 90376, + "tool_calls": 7, + "turns": 8 + }, + "world-047": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 22661, + "fixed_tokens": 1109, + "processed_tokens": 125256, + "tool_calls": 7, + "turns": 8 + }, + "world-048": { + "answer": "dec-compile-gull-to-bytecode-for-a-register-vm--jrqje646fpevp5f3", + "tokens": 37121, + "fixed_tokens": 1096, + "processed_tokens": 725767, + "tool_calls": 31, + "turns": 32 + }, + "world-049": { + "answer": "22%", + "tokens": 33472, + "fixed_tokens": 1108, + "processed_tokens": 561531, + "tool_calls": 21, + "turns": 22 + }, + "world-050": { + "answer": "fnd-alpha-content-needs-6-2-ms-of-gull-time-per-fram--4mekgdpsdv5nk085", + "tokens": 26311, + "fixed_tokens": 1103, + "processed_tokens": 336536, + "tool_calls": 15, + "turns": 16 + }, + "world-051": { + "answer": "The tide simulation moved to 32.32 fixed point so that tide state is bit-identical on every CPU and compiler; float tides drifted apart between machines within 6 minutes, and integer math avoids that where strict IEEE mode would not. The strict IEEE option was rejected in favor of it (`dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e`).", + "tokens": 36525, + "fixed_tokens": 1105, + "processed_tokens": 659715, + "tool_calls": 23, + "turns": 24 + }, + "world-052": { + "answer": "Dara Whitlock", + "tokens": 24288, + "fixed_tokens": 1104, + "processed_tokens": 231939, + "tool_calls": 11, + "turns": 12 + }, + "world-053": { + "answer": "unknown", + "tokens": 92958, + "fixed_tokens": 1102, + "processed_tokens": 2383883, + "tool_calls": 40, + "turns": 39 + }, + "world-054": { + "answer": "Marta Lindqvist", + "tokens": 28512, + "fixed_tokens": 1105, + "processed_tokens": 521729, + "tool_calls": 22, + "turns": 23 + }, + "world-055": { + "answer": "Kelp stays because dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew found that Kelp with a 1 MB dictionary still beats Zstandard cold on Skiff, and it closed the reopened Zstandard option. That decision's 1.7x figure is out of date: the later finding fnd-kelp-with-a-1-mb-dictionary-decodes-at-1-5-gb-s--y9v8jrxgbv6abrx5 puts Kelp's lead at 1.25x.", + "tokens": 60627, + "fixed_tokens": 1099, + "processed_tokens": 1024843, + "tool_calls": 24, + "turns": 25 + }, + "world-056": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 23989, + "fixed_tokens": 1103, + "processed_tokens": 197302, + "tool_calls": 10, + "turns": 11 + }, + "world-057": { + "answer": "tile [212, 87]", + "tokens": 20722, + "fixed_tokens": 1104, + "processed_tokens": 255438, + "tool_calls": 15, + "turns": 16 + }, + "world-058": { + "answer": "0.8 GB/s", + "tokens": 35754, + "fixed_tokens": 1101, + "processed_tokens": 531102, + "tool_calls": 19, + "turns": 20 + }, + "world-059": { + "answer": "dec-derive-the-missing-harmonic-phase-from-each-save--gbqrvvgyq2gy2gnw", + "tokens": 13179, + "fixed_tokens": 1106, + "processed_tokens": 136554, + "tool_calls": 10, + "turns": 11 + }, + "world-060": { + "answer": "dec-partition-the-world-into-192-m-cells--bk9r1c5jwb6fh8tc", + "tokens": 19699, + "fixed_tokens": 1106, + "processed_tokens": 262693, + "tool_calls": 15, + "turns": 16 + }, + "world-061": { + "answer": "236 MB", + "tokens": 30625, + "fixed_tokens": 1104, + "processed_tokens": 520068, + "tool_calls": 21, + "turns": 21 + }, + "world-062": { + "answer": "310 MB", + "tokens": 19333, + "fixed_tokens": 1108, + "processed_tokens": 324032, + "tool_calls": 19, + "turns": 20 + }, + "world-063": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 27564, + "fixed_tokens": 1103, + "processed_tokens": 458558, + "tool_calls": 19, + "turns": 20 + }, + "world-064": { + "answer": [ + "dec-force-strict-ieee-float-mode-for-the-tide-simula--ecwmqt7ncjngysqw", + "dec-have-the-host-stream-tide-heights-to-every-peer--a3rb1seedn2bj7k8" + ], + "tokens": 35659, + "fixed_tokens": 1101, + "processed_tokens": 666286, + "tool_calls": 27, + "turns": 27 + }, + "world-065": { + "answer": "3.1 ms", + "tokens": 35511, + "fixed_tokens": 1108, + "processed_tokens": 913217, + "tool_calls": 36, + "turns": 37 + }, + "world-066": { + "answer": "46", + "tokens": 37075, + "fixed_tokens": 1106, + "processed_tokens": 612052, + "tool_calls": 21, + "turns": 22 + }, + "world-067": { + "answer": "1.2 GB/s", + "tokens": 24987, + "fixed_tokens": 1104, + "processed_tokens": 331309, + "tool_calls": 21, + "turns": 17 + }, + "world-068": { + "answer": true, + "tokens": 45833, + "fixed_tokens": 1103, + "processed_tokens": 500103, + "tool_calls": 30, + "turns": 19 + }, + "world-069": { + "answer": 6, + "tokens": 63925, + "fixed_tokens": 1101, + "processed_tokens": 590814, + "tool_calls": 19, + "turns": 18 + }, + "world-070": { + "answer": 3, + "tokens": 8448, + "fixed_tokens": 1100, + "processed_tokens": 74254, + "tool_calls": 7, + "turns": 8 + }, + "world-071": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--yxsjd26wcaz6fgx9" + ], + "tokens": 40587, + "fixed_tokens": 1098, + "processed_tokens": 414503, + "tool_calls": 15, + "turns": 16 + }, + "world-072": { + "answer": [ + "dec-compress-streamed-assets-with-zstandard--bs1t15cygc51cpje", + "dec-compress-streamed-assets-with-lz4--v1pfsgfp0b3yjv4s" + ], + "tokens": 68205, + "fixed_tokens": 1103, + "processed_tokens": 1972310, + "tool_calls": 46, + "turns": 47 + }, + "world-073": { + "answer": false, + "tokens": 58411, + "fixed_tokens": 1108, + "processed_tokens": 727508, + "tool_calls": 19, + "turns": 20 + }, + "world-074": { + "answer": 7, + "tokens": 9115, + "fixed_tokens": 1103, + "processed_tokens": 69149, + "tool_calls": 6, + "turns": 7 + }, + "world-075": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--8679sjn071ycsbpv", + "dec-partition-the-world-into-192-meter-cells--q62ddwdv2n8wxay6", + "dec-return-engine-jobs-to-the-thread-pool-with-conti--547qbweqvrp1j5g3" + ], + "tokens": 7021, + "fixed_tokens": 1102, + "processed_tokens": 60801, + "tool_calls": 6, + "turns": 7 + }, + "world-076": { + "answer": "unknown", + "tokens": 101576, + "fixed_tokens": 1108, + "processed_tokens": 5864116, + "tool_calls": 91, + "turns": 92 + }, + "world-077": { + "answer": 251, + "tokens": 31130, + "fixed_tokens": 1100, + "processed_tokens": 443606, + "tool_calls": 19, + "turns": 20 + }, + "world-078": { + "answer": [ + "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "rsk-reworking-the-read-scheduler-may-slip-the-coast--9jzvs7etvb5jxqtp" + ], + "tokens": 10229, + "fixed_tokens": 1107, + "processed_tokens": 57628, + "tool_calls": 5, + "turns": 6 + }, + "world-079": { + "answer": [ + "con-streaming-pool-budget--8czwb4fjr5p5vpy7", + "con-no-runtime-code-generation-on-consoles--6cf25zqmsvqkdd73", + "con-tide-state-must-be-bit-identical-on-every-peer--zrpnp8qjyxzkas7t", + "con-saves-from-beta-onward-must-load-in-every-later--v6z856sxtgc5sb09" + ], + "tokens": 12852, + "fixed_tokens": 1103, + "processed_tokens": 84647, + "tool_calls": 7, + "turns": 8 + }, + "world-080": { + "answer": [ + "iss-gull-hot-reload-rejects-struct-layout-changes--3v2w2f9hg2gwy6aj", + "iss-gull-hot-reload-drops-suspended-coroutines--z7j2gmyny6gdxd07", + "iss-cross-platform-cloud-saves-are-unplanned--5bk0k4wrgj42ztfy" + ], + "tokens": 39751, + "fixed_tokens": 1100, + "processed_tokens": 435208, + "tool_calls": 16, + "turns": 17 + }, + "world-081": { + "answer": "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "tokens": 40329, + "fixed_tokens": 1102, + "processed_tokens": 812673, + "tool_calls": 27, + "turns": 28 + } + } +} diff --git a/benchmarks/proof-recall/runs/x10-find-r2-find.json b/benchmarks/proof-recall/runs/x10-find-r2-find.json new file mode 100644 index 00000000..4aaba7a5 --- /dev/null +++ b/benchmarks/proof-recall/runs/x10-find-r2-find.json @@ -0,0 +1,2259 @@ +{ + "run": { + "name": "x10-find-r2-find", + "protocol": "one fresh session per question", + "models": ["claude-haiku-5-5"], + "find_hash": "e5be653299abe9b2329822d98ffe09e90ee7cfe96f8711c813a51dc6500b887f" + }, + "answers": { + "ambiguous-001": { + "answer": "384 MB", + "tokens": 11473, + "fixed_tokens": 1103, + "processed_tokens": 79614, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-002": { + "answer": "99.6 percent", + "tokens": 7729, + "fixed_tokens": 1109, + "processed_tokens": 58603, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-003": { + "answer": "92.4%", + "tokens": 9905, + "fixed_tokens": 1104, + "processed_tokens": 87031, + "tool_calls": 7, + "turns": 8 + }, + "ambiguous-004": { + "answer": "6 min 40 s", + "tokens": 8248, + "fixed_tokens": 1104, + "processed_tokens": 59976, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-005": { + "answer": "6 weeks", + "tokens": 8086, + "fixed_tokens": 1103, + "processed_tokens": 59605, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-006": { + "answer": "16-bit float per channel", + "tokens": 21663, + "fixed_tokens": 1103, + "processed_tokens": 158532, + "tool_calls": 9, + "turns": 9 + }, + "ambiguous-007": { + "answer": "32 steps", + "tokens": 13363, + "fixed_tokens": 1102, + "processed_tokens": 81976, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-008": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 10192, + "fixed_tokens": 1102, + "processed_tokens": 88008, + "tool_calls": 7, + "turns": 8 + }, + "ambiguous-009": { + "answer": "3.0 ms", + "tokens": 11132, + "fixed_tokens": 1106, + "processed_tokens": 77181, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-010": { + "answer": "160 MB/s", + "tokens": 9384, + "fixed_tokens": 1102, + "processed_tokens": 62343, + "tool_calls": 7, + "turns": 6 + }, + "ambiguous-011": { + "answer": "260 MB", + "tokens": 11330, + "fixed_tokens": 1103, + "processed_tokens": 91500, + "tool_calls": 9, + "turns": 8 + }, + "ambiguous-012": { + "answer": "11 m", + "tokens": 11036, + "fixed_tokens": 1102, + "processed_tokens": 82102, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-013": { + "answer": "61%", + "tokens": 8036, + "fixed_tokens": 1103, + "processed_tokens": 59928, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-014": { + "answer": "textures", + "tokens": 8020, + "fixed_tokens": 1103, + "processed_tokens": 59388, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-015": { + "answer": "unknown", + "tokens": 28847, + "fixed_tokens": 1108, + "processed_tokens": 239207, + "tool_calls": 17, + "turns": 12 + }, + "ambiguous-016": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 9406, + "fixed_tokens": 1104, + "processed_tokens": 61872, + "tool_calls": 7, + "turns": 6 + }, + "ambiguous-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 10732, + "fixed_tokens": 1104, + "processed_tokens": 118605, + "tool_calls": 9, + "turns": 10 + }, + "ambiguous-018": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 8034, + "fixed_tokens": 1099, + "processed_tokens": 59921, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-019": { + "answer": "7.4 ms", + "tokens": 7946, + "fixed_tokens": 1106, + "processed_tokens": 59442, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-020": { + "answer": "5912", + "tokens": 7991, + "fixed_tokens": 1111, + "processed_tokens": 59411, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-021": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 7999, + "fixed_tokens": 1107, + "processed_tokens": 59544, + "tool_calls": 5, + "turns": 6 + }, + "ambiguous-022": { + "answer": "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "tokens": 14191, + "fixed_tokens": 1111, + "processed_tokens": 105937, + "tool_calls": 7, + "turns": 8 + }, + "ambiguous-023": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 8182, + "fixed_tokens": 1105, + "processed_tokens": 68844, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-024": { + "answer": "fnd-streaming-tide-heights-costs-46-kb-s-per-peer--knxqtyqh6xcz1tp9", + "tokens": 13172, + "fixed_tokens": 1105, + "processed_tokens": 87617, + "tool_calls": 8, + "turns": 7 + }, + "cross-001": { + "answer": "con-frame-budget--br0qszfg8fys908n", + "tokens": 7609, + "fixed_tokens": 1104, + "processed_tokens": 58873, + "tool_calls": 5, + "turns": 6 + }, + "cross-002": { + "answer": 2, + "tokens": 14922, + "fixed_tokens": 1091, + "processed_tokens": 95363, + "tool_calls": 8, + "turns": 7 + }, + "cross-003": { + "answer": 2, + "tokens": 5390, + "fixed_tokens": 1097, + "processed_tokens": 32150, + "tool_calls": 3, + "turns": 4 + }, + "cross-004": { + "answer": "dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e", + "tokens": 11250, + "fixed_tokens": 1096, + "processed_tokens": 77908, + "tool_calls": 6, + "turns": 7 + }, + "cross-005": { + "answer": "Rhea Valdés", + "tokens": 14207, + "fixed_tokens": 1094, + "processed_tokens": 128302, + "tool_calls": 9, + "turns": 10 + }, + "cross-006": { + "answer": "~19%", + "tokens": 9574, + "fixed_tokens": 1093, + "processed_tokens": 63044, + "tool_calls": 7, + "turns": 6 + }, + "cross-007": { + "answer": "8 GB", + "tokens": 10612, + "fixed_tokens": 1096, + "processed_tokens": 117627, + "tool_calls": 9, + "turns": 10 + }, + "cross-008": { + "answer": "9 GB", + "tokens": 16976, + "fixed_tokens": 1103, + "processed_tokens": 150833, + "tool_calls": 12, + "turns": 11 + }, + "cross-009": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 13974, + "fixed_tokens": 1096, + "processed_tokens": 99998, + "tool_calls": 9, + "turns": 8 + }, + "cross-010": { + "answer": "dec-build-an-on-device-pipeline-cache-for-skiff--726r9mtk76qndeh7", + "tokens": 13298, + "fixed_tokens": 1093, + "processed_tokens": 87291, + "tool_calls": 8, + "turns": 7 + }, + "cross-011": { + "answer": false, + "tokens": 10049, + "fixed_tokens": 1094, + "processed_tokens": 87525, + "tool_calls": 7, + "turns": 8 + }, + "cross-012": { + "answer": "61 MB/s", + "tokens": 7860, + "fixed_tokens": 1100, + "processed_tokens": 58865, + "tool_calls": 5, + "turns": 6 + }, + "cross-013": { + "answer": "Skiff runs Tern at the Performance preset, saving 1.8 ms over Quality (3.9 ms vs 2.1 ms), per dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3.", + "tokens": 13546, + "fixed_tokens": 1101, + "processed_tokens": 115560, + "tool_calls": 8, + "turns": 9 + }, + "cross-014": { + "answer": 3, + "tokens": 4792, + "fixed_tokens": 1093, + "processed_tokens": 39237, + "tool_calls": 4, + "turns": 5 + }, + "cross-015": { + "answer": "Two Efforts are no longer active. The co-op rollback netcode prototype was abandoned: dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b stopped it because re-running tide simulation on every rollback frame costs 4.8 ms on Skiff. The shader compile cache and build farm was completed once the cache ran on every farm build (dec-turn-the-shader-compile-cache-on-for-every-farm--4rr3b12b4sasdmaj), as its retrospective fnd-retrospective-on-the-shader-compile-cache--9rpv2a60ach04cx6 records, with clean rebuilds down from 47 minutes to 6 min 40 s.", + "tokens": 38571, + "fixed_tokens": 1097, + "processed_tokens": 242711, + "tool_calls": 13, + "turns": 10 + }, + "cross-016": { + "answer": "16.7 ms", + "tokens": 13012, + "fixed_tokens": 1095, + "processed_tokens": 83564, + "tool_calls": 8, + "turns": 7 + }, + "cross-017": { + "answer": "unknown", + "tokens": 14539, + "fixed_tokens": 1095, + "processed_tokens": 97176, + "tool_calls": 7, + "turns": 8 + }, + "netcode-001": { + "answer": "abandoned", + "tokens": 7683, + "fixed_tokens": 1096, + "processed_tokens": 59106, + "tool_calls": 5, + "turns": 6 + }, + "netcode-002": { + "answer": "netcode/rollback-archive", + "tokens": 7887, + "fixed_tokens": 1102, + "processed_tokens": 58951, + "tool_calls": 5, + "turns": 6 + }, + "netcode-003": { + "answer": "38 KB", + "tokens": 10175, + "fixed_tokens": 1100, + "processed_tokens": 89096, + "tool_calls": 7, + "turns": 8 + }, + "netcode-004": { + "answer": "The team dropped rollback because each rollback frame must re-run the tide simulation, costing 4.8 ms of a 25 ms frame on Skiff, and quarter-resolution re-simulation desynced within 90 seconds. The reason depends on the world Decision dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e, which keeps tides deterministic and which co-op needs.", + "tokens": 11805, + "fixed_tokens": 1100, + "processed_tokens": 78319, + "tool_calls": 6, + "turns": 7 + }, + "netcode-005": { + "answer": "unknown", + "tokens": 13429, + "fixed_tokens": 1097, + "processed_tokens": 81910, + "tool_calls": 6, + "turns": 7 + }, + "netcode-006": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 9978, + "fixed_tokens": 1097, + "processed_tokens": 86458, + "tool_calls": 7, + "turns": 8 + }, + "netcode-007": { + "answer": "dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b", + "tokens": 8102, + "fixed_tokens": 1096, + "processed_tokens": 59681, + "tool_calls": 5, + "turns": 6 + }, + "netcode-008": { + "answer": "dec-use-a-4-frame-rollback-window--pqe4hvbgmbp0dmh9", + "tokens": 10014, + "fixed_tokens": 1099, + "processed_tokens": 88488, + "tool_calls": 7, + "turns": 8 + }, + "netcode-009": { + "answer": "1.9 ms", + "tokens": 9155, + "fixed_tokens": 1100, + "processed_tokens": 86362, + "tool_calls": 7, + "turns": 8 + }, + "netcode-010": { + "answer": "dec-use-a-6-frame-rollback-window-with-1-frame-of-in--s53944zhkw103frk", + "tokens": 11477, + "fixed_tokens": 1105, + "processed_tokens": 75745, + "tool_calls": 6, + "turns": 7 + }, + "netcode-011": { + "answer": "con-co-op-peers-must-agree-on-tide-state-bit-for-bit--4v0sdbc3t4ywzr3g", + "tokens": 24102, + "fixed_tokens": 1094, + "processed_tokens": 245001, + "tool_calls": 15, + "turns": 14 + }, + "netcode-012": { + "answer": "dec-re-simulate-tides-at-quarter-resolution-during-r--7ah25p0r0xt9719f", + "tokens": 20208, + "fixed_tokens": 1099, + "processed_tokens": 154468, + "tool_calls": 10, + "turns": 11 + }, + "netcode-013": { + "answer": [ + "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "fnd-rollback-re-simulates-a-4-frame-window-in-1-1-ms--xa30a1cez2brphyf", + "rsk-re-simulation-cost-could-blow-the-skiff-frame-bu--sk2njsp170a3qwyf" + ], + "tokens": 21853, + "fixed_tokens": 1097, + "processed_tokens": 161848, + "tool_calls": 21, + "turns": 10 + }, + "netcode-014": { + "answer": [ + "iss-late-join-during-a-tide-surge-desyncs-the-joinin--mya93va0n7s8g7cj" + ], + "tokens": 35469, + "fixed_tokens": 1100, + "processed_tokens": 453101, + "tool_calls": 17, + "turns": 18 + }, + "netcode-015": { + "answer": 2, + "tokens": 11064, + "fixed_tokens": 1098, + "processed_tokens": 103602, + "tool_calls": 9, + "turns": 10 + }, + "ports-001": { + "answer": "25 ms", + "tokens": 7636, + "fixed_tokens": 1102, + "processed_tokens": 58620, + "tool_calls": 5, + "turns": 6 + }, + "ports-002": { + "answer": "384 MB", + "tokens": 8196, + "fixed_tokens": 1096, + "processed_tokens": 60037, + "tool_calls": 5, + "turns": 6 + }, + "ports-003": { + "answer": "61 MB/s", + "tokens": 9345, + "fixed_tokens": 1098, + "processed_tokens": 87455, + "tool_calls": 7, + "turns": 8 + }, + "ports-004": { + "answer": "accepted", + "tokens": 7326, + "fixed_tokens": 1095, + "processed_tokens": 57690, + "tool_calls": 5, + "turns": 6 + }, + "ports-005": { + "answer": "unknown", + "tokens": 19887, + "fixed_tokens": 1098, + "processed_tokens": 185113, + "tool_calls": 10, + "turns": 11 + }, + "ports-034": { + "answer": "6 weeks", + "tokens": 7892, + "fixed_tokens": 1101, + "processed_tokens": 59253, + "tool_calls": 5, + "turns": 6 + }, + "ports-006": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 9029, + "fixed_tokens": 1094, + "processed_tokens": 61664, + "tool_calls": 7, + "turns": 6 + }, + "ports-007": { + "answer": "SKF-UI-031", + "tokens": 7475, + "fixed_tokens": 1099, + "processed_tokens": 58040, + "tool_calls": 5, + "turns": 6 + }, + "ports-008": { + "answer": "dec-use-a-spatial-upscaler-with-sharpening-on-skiff--cynznx2wg9hswzgb", + "tokens": 8092, + "fixed_tokens": 1104, + "processed_tokens": 59377, + "tool_calls": 5, + "turns": 6 + }, + "ports-009": { + "answer": "dec-load-skiff-fast-travel-without-a-loading-screen--v4qft5znysjv5f24", + "tokens": 7882, + "fixed_tokens": 1097, + "processed_tokens": 59207, + "tool_calls": 5, + "turns": 6 + }, + "ports-010": { + "answer": "con-frame-budget--wt5rfqh5z444g2cg", + "tokens": 14854, + "fixed_tokens": 1098, + "processed_tokens": 131887, + "tool_calls": 11, + "turns": 10 + }, + "ports-011": { + "answer": "99.6 percent", + "tokens": 7607, + "fixed_tokens": 1094, + "processed_tokens": 58360, + "tool_calls": 5, + "turns": 6 + }, + "ports-012": { + "answer": "The team raised the floor to 576p because caustic patterns under saltglass shimmered badly at the 540p floor even with Tern (dec-raise-the-skiff-dynamic-resolution-floor-to-576p--fep9q7m08pp9f0bb, derived from fnd-saltglass-caustics-shimmer-at-the-540p-floor--0aa8sts89vhq1s6q).", + "tokens": 8223, + "fixed_tokens": 1092, + "processed_tokens": 59086, + "tool_calls": 5, + "turns": 6 + }, + "ports-013": { + "answer": "Ship Tern as Lantern's only upscaler", + "tokens": 8614, + "fixed_tokens": 1100, + "processed_tokens": 61166, + "tool_calls": 5, + "turns": 6 + }, + "ports-014": { + "answer": "dec-scale-skiff-resolution-dynamically-between-540p--jtytvgh2erd32x68", + "tokens": 17572, + "fixed_tokens": 1096, + "processed_tokens": 148552, + "tool_calls": 9, + "turns": 10 + }, + "ports-015": { + "answer": "textures", + "tokens": 10003, + "fixed_tokens": 1095, + "processed_tokens": 89094, + "tool_calls": 7, + "turns": 8 + }, + "ports-016": { + "answer": "dec-ship-prebuilt-pipeline-blobs-in-the-skiff-packag--3vjysjgy9qkvzebn", + "tokens": 13998, + "fixed_tokens": 1103, + "processed_tokens": 127301, + "tool_calls": 9, + "turns": 10 + }, + "ports-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 8359, + "fixed_tokens": 1094, + "processed_tokens": 60491, + "tool_calls": 5, + "turns": 6 + }, + "ports-018": { + "answer": "504p", + "tokens": 8818, + "fixed_tokens": 1097, + "processed_tokens": 70577, + "tool_calls": 6, + "turns": 7 + }, + "ports-019": { + "answer": "4 MB per minute", + "tokens": 8572, + "fixed_tokens": 1093, + "processed_tokens": 83727, + "tool_calls": 7, + "turns": 8 + }, + "ports-020": { + "answer": "fnd-skiff-sdk-1-2-takes-1-gb-back-from-games--a0ybdgjzw64hwfqa", + "tokens": 11123, + "fixed_tokens": 1096, + "processed_tokens": 119642, + "tool_calls": 9, + "turns": 10 + }, + "ports-021": { + "answer": "SKF-DSP-004", + "tokens": 9501, + "fixed_tokens": 1100, + "processed_tokens": 94682, + "tool_calls": 8, + "turns": 9 + }, + "ports-022": { + "answer": "cit-skiff-sdk-1-2-release-notes--7gwwn8xxffrx6npe", + "tokens": 10142, + "fixed_tokens": 1095, + "processed_tokens": 116156, + "tool_calls": 9, + "turns": 10 + }, + "ports-023": { + "answer": "fnd-tern-costs-3-9-ms-at-quality-and-2-1-ms-at-perfo--dkbskbqfgfwtryq0", + "tokens": 14697, + "fixed_tokens": 1096, + "processed_tokens": 147940, + "tool_calls": 10, + "turns": 11 + }, + "ports-024": { + "answer": "fnd-skiff-firmware-1-2-rejects-pipeline-blobs-built--pernvgb2he7m3z3y", + "tokens": 11271, + "fixed_tokens": 1093, + "processed_tokens": 118174, + "tool_calls": 9, + "turns": 10 + }, + "ports-025": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 9801, + "fixed_tokens": 1101, + "processed_tokens": 88647, + "tool_calls": 7, + "turns": 8 + }, + "ports-026": { + "answer": "912 MB", + "tokens": 12770, + "fixed_tokens": 1097, + "processed_tokens": 154801, + "tool_calls": 11, + "turns": 12 + }, + "ports-027": { + "answer": [ + "dec-stream-skiff-world-tiles-in-2-mb-reads--3jz9w9gvhhc7eppw", + "rsk-tide-streaming-outruns-skiff-storage--6q8s3b383z2tyeyb" + ], + "tokens": 53308, + "fixed_tokens": 1097, + "processed_tokens": 549241, + "tool_calls": 26, + "turns": 20 + }, + "ports-028": { + "answer": 2, + "tokens": 9488, + "fixed_tokens": 1094, + "processed_tokens": 86448, + "tool_calls": 8, + "turns": 9 + }, + "ports-029": { + "answer": 8, + "tokens": 74444, + "fixed_tokens": 1095, + "processed_tokens": 1324230, + "tool_calls": 30, + "turns": 31 + }, + "ports-030": { + "answer": [ + "dec-cap-tide-storm-particles-at-6-000-on-skiff--resze52afjxyxys1", + "dec-drop-skiff-to-30-hz-while-a-tide-storm-is-on-scr--72vx1c8s4jrs6axc" + ], + "tokens": 9159, + "fixed_tokens": 1094, + "processed_tokens": 61745, + "tool_calls": 5, + "turns": 6 + }, + "ports-031": { + "answer": [ + "iss-meridian-cert-activity-cards-missing-for-two-bio--k214cmfra7pyvvzq", + "iss-skiff-cert-docked-30-hz-judder-on-some-tvs--9nq2xg9mzash474g" + ], + "tokens": 37105, + "fixed_tokens": 1094, + "processed_tokens": 381673, + "tool_calls": 16, + "turns": 15 + }, + "ports-032": { + "answer": 2, + "tokens": 7035, + "fixed_tokens": 1099, + "processed_tokens": 63121, + "tool_calls": 6, + "turns": 7 + }, + "ports-033": { + "answer": "unknown", + "tokens": 9949, + "fixed_tokens": 1096, + "processed_tokens": 60767, + "tool_calls": 5, + "turns": 6 + }, + "render-001": { + "answer": "7.4 ms", + "tokens": 8007, + "fixed_tokens": 1094, + "processed_tokens": 59686, + "tool_calls": 5, + "turns": 6 + }, + "render-002": { + "answer": "3.4x", + "tokens": 8859, + "fixed_tokens": 1093, + "processed_tokens": 84068, + "tool_calls": 7, + "turns": 8 + }, + "render-003": { + "answer": "Tomasz Brandt", + "tokens": 9002, + "fixed_tokens": 1092, + "processed_tokens": 86057, + "tool_calls": 7, + "turns": 8 + }, + "render-004": { + "answer": "rejected", + "tokens": 7810, + "fixed_tokens": 1095, + "processed_tokens": 58791, + "tool_calls": 5, + "turns": 6 + }, + "render-005": { + "answer": true, + "tokens": 7605, + "fixed_tokens": 1088, + "processed_tokens": 57991, + "tool_calls": 5, + "turns": 6 + }, + "render-006": { + "answer": "unknown", + "tokens": 11354, + "fixed_tokens": 1090, + "processed_tokens": 92413, + "tool_calls": 7, + "turns": 8 + }, + "render-007": { + "answer": "2026-07-24", + "tokens": 11570, + "fixed_tokens": 1097, + "processed_tokens": 79456, + "tool_calls": 6, + "turns": 7 + }, + "render-008": { + "answer": "2 failures", + "tokens": 7569, + "fixed_tokens": 1097, + "processed_tokens": 57950, + "tool_calls": 5, + "turns": 6 + }, + "render-009": { + "answer": "32 steps", + "tokens": 7817, + "fixed_tokens": 1093, + "processed_tokens": 58649, + "tool_calls": 5, + "turns": 6 + }, + "render-010": { + "answer": "unknown", + "tokens": 10613, + "fixed_tokens": 1090, + "processed_tokens": 88900, + "tool_calls": 7, + "turns": 8 + }, + "render-011": { + "answer": "fnd-sdf-gi-costs-2-4-ms-on-meridian--2w6c2pc5zmcsyjre", + "tokens": 8228, + "fixed_tokens": 1097, + "processed_tokens": 60374, + "tool_calls": 5, + "turns": 6 + }, + "render-012": { + "answer": "deferred", + "tokens": 9352, + "fixed_tokens": 1095, + "processed_tokens": 87062, + "tool_calls": 7, + "turns": 8 + }, + "render-013": { + "answer": "16-bit float per channel", + "tokens": 11474, + "fixed_tokens": 1092, + "processed_tokens": 78592, + "tool_calls": 6, + "turns": 7 + }, + "render-014": { + "answer": "52%", + "tokens": 8287, + "fixed_tokens": 1088, + "processed_tokens": 60265, + "tool_calls": 5, + "turns": 6 + }, + "render-015": { + "answer": "4 minutes", + "tokens": 7910, + "fixed_tokens": 1089, + "processed_tokens": 59253, + "tool_calls": 5, + "turns": 6 + }, + "render-016": { + "answer": "1.2 km", + "tokens": 8138, + "fixed_tokens": 1089, + "processed_tokens": 60023, + "tool_calls": 5, + "turns": 6 + }, + "render-017": { + "answer": "unknown", + "tokens": 13718, + "fixed_tokens": 1093, + "processed_tokens": 97485, + "tool_calls": 7, + "turns": 8 + }, + "render-018": { + "answer": "unknown", + "tokens": 13935, + "fixed_tokens": 1095, + "processed_tokens": 82590, + "tool_calls": 6, + "turns": 7 + }, + "render-019": { + "answer": "dec-use-the-custom-brine-curve-for-tonemapping--p735g6wf2njare86", + "tokens": 7777, + "fixed_tokens": 1093, + "processed_tokens": 58936, + "tool_calls": 5, + "turns": 6 + }, + "render-020": { + "answer": "fnd-screen-space-refraction-loses-off-screen-content--k8zvmmdee7wg2fjy", + "tokens": 7470, + "fixed_tokens": 1096, + "processed_tokens": 57882, + "tool_calls": 5, + "turns": 6 + }, + "render-021": { + "answer": "dec-adopt-a-visibility-buffer-for-all-opaque-geometr--68vchzqgd1avw8q4", + "tokens": 10563, + "fixed_tokens": 1087, + "processed_tokens": 89065, + "tool_calls": 7, + "turns": 8 + }, + "render-022": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 13071, + "fixed_tokens": 1092, + "processed_tokens": 112641, + "tool_calls": 8, + "turns": 9 + }, + "render-023": { + "answer": "dec-capture-the-cliffs-and-dunes-benchmarks-on-every--3pqyr66yh4v6q676", + "tokens": 7035, + "fixed_tokens": 1097, + "processed_tokens": 56312, + "tool_calls": 5, + "turns": 6 + }, + "render-024": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 7791, + "fixed_tokens": 1096, + "processed_tokens": 59267, + "tool_calls": 5, + "turns": 6 + }, + "render-025": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-after-fp16-pa--p4gcmzw2a4bvzjev", + "tokens": 19830, + "fixed_tokens": 1100, + "processed_tokens": 122311, + "tool_calls": 11, + "turns": 8 + }, + "render-026": { + "answer": "dec-ship-four-tern-quality-presets--vjw18d82dx5gath8", + "tokens": 7443, + "fixed_tokens": 1095, + "processed_tokens": 57942, + "tool_calls": 5, + "turns": 6 + }, + "render-027": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 11402, + "fixed_tokens": 1091, + "processed_tokens": 76296, + "tool_calls": 10, + "turns": 7 + }, + "render-028": { + "answer": "dec-move-to-gpu-driven-two-phase-occlusion-culling--6dre481mabxsw6ee", + "tokens": 7657, + "fixed_tokens": 1099, + "processed_tokens": 58037, + "tool_calls": 5, + "turns": 6 + }, + "render-029": { + "answer": "fnd-vsm-page-pool-overflows-during-spring-tides--m2p8e89hhdmh0z8a", + "tokens": 7948, + "fixed_tokens": 1091, + "processed_tokens": 59845, + "tool_calls": 5, + "turns": 6 + }, + "render-030": { + "answer": "unknown", + "tokens": 12846, + "fixed_tokens": 1091, + "processed_tokens": 105338, + "tool_calls": 8, + "turns": 9 + }, + "render-031": { + "answer": "The third version, dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9, moved foliage and saltglass out of the buffer to render them in a forward pass after the resolve: alpha-tested foliage cost 2.1 ms in the buffer versus 0.8 ms in forward, and saltglass refracts the resolved scene, so it needs that scene to exist first.", + "tokens": 8415, + "fixed_tokens": 1095, + "processed_tokens": 59579, + "tool_calls": 5, + "turns": 6 + }, + "render-032": { + "answer": "Tern was chosen as Lantern's only upscaler (decision dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e): Halcyon has no Skiff build, and shipping two upscalers would break the one-upscaler constraint. Owning the code also mattered more than raw quality for saltglass, since the team can change Tern upscaling the same week without a vendor ticket or license.", + "tokens": 8012, + "fixed_tokens": 1092, + "processed_tokens": 58816, + "tool_calls": 5, + "turns": 6 + }, + "render-033": { + "answer": "128 MB", + "tokens": 8199, + "fixed_tokens": 1092, + "processed_tokens": 60424, + "tool_calls": 5, + "turns": 6 + }, + "render-034": { + "answer": "PC workstation (Meridian clock profile)", + "tokens": 9865, + "fixed_tokens": 1093, + "processed_tokens": 89517, + "tool_calls": 7, + "turns": 8 + }, + "render-035": { + "answer": "5912 pages", + "tokens": 9535, + "fixed_tokens": 1095, + "processed_tokens": 87923, + "tool_calls": 7, + "turns": 8 + }, + "render-036": { + "answer": "HDR-07", + "tokens": 7554, + "fixed_tokens": 1101, + "processed_tokens": 58247, + "tool_calls": 5, + "turns": 6 + }, + "render-037": { + "answer": "0.33", + "tokens": 11831, + "fixed_tokens": 1097, + "processed_tokens": 148033, + "tool_calls": 11, + "turns": 12 + }, + "render-038": { + "answer": "dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9", + "tokens": 21063, + "fixed_tokens": 1090, + "processed_tokens": 188686, + "tool_calls": 10, + "turns": 11 + }, + "render-039": { + "answer": "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "tokens": 7895, + "fixed_tokens": 1089, + "processed_tokens": 77528, + "tool_calls": 7, + "turns": 8 + }, + "render-040": { + "answer": "fnd-lantern-at-alpha--ckpd9ecvdjqbcr7c", + "tokens": 11394, + "fixed_tokens": 1097, + "processed_tokens": 95297, + "tool_calls": 8, + "turns": 9 + }, + "render-041": { + "answer": "4.7 ms", + "tokens": 9874, + "fixed_tokens": 1092, + "processed_tokens": 64168, + "tool_calls": 7, + "turns": 6 + }, + "render-042": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 8435, + "fixed_tokens": 1090, + "processed_tokens": 60646, + "tool_calls": 5, + "turns": 6 + }, + "render-043": { + "answer": "fnd-meshlet-cone-culling-rejects-9-percent-of-triang--t6e5x7trzedex105", + "tokens": 7795, + "fixed_tokens": 1091, + "processed_tokens": 58463, + "tool_calls": 5, + "turns": 6 + }, + "render-044": { + "answer": "dec-screen-space-refraction-with-a-ray-marched-fallb--2639xtpyrfrvbx1r", + "tokens": 7775, + "fixed_tokens": 1098, + "processed_tokens": 58640, + "tool_calls": 5, + "turns": 6 + }, + "render-045": { + "answer": "false", + "tokens": 11256, + "fixed_tokens": 1091, + "processed_tokens": 82324, + "tool_calls": 6, + "turns": 7 + }, + "render-046": { + "answer": "Reopened on 2026-01-12 because the 3.1 ms cost finding that rejected it was a capture error; with cached pages, virtual shadow maps cost 1.4 ms, less than four cascades (dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4).", + "tokens": 8646, + "fixed_tokens": 1092, + "processed_tokens": 60738, + "tool_calls": 5, + "turns": 6 + }, + "render-047": { + "answer": "Because the release-build capture put the visibility buffer on Skiff at 7.4 ms against 8.1 ms for forward+ (6.3 ms with half-rate material resolve), so keeping the forward+ fallback was no longer worth it, and maintaining two opaque paths cost about one engineer-week a month. Record: dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 12346, + "fixed_tokens": 1091, + "processed_tokens": 93209, + "tool_calls": 7, + "turns": 8 + }, + "render-048": { + "answer": "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "tokens": 15841, + "fixed_tokens": 1093, + "processed_tokens": 129995, + "tool_calls": 9, + "turns": 10 + }, + "render-049": { + "answer": "dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4", + "tokens": 9563, + "fixed_tokens": 1097, + "processed_tokens": 85552, + "tool_calls": 7, + "turns": 8 + }, + "render-050": { + "answer": "con-frame-budget--fzdrbxns6dmrrmy3", + "tokens": 12766, + "fixed_tokens": 1092, + "processed_tokens": 86594, + "tool_calls": 7, + "turns": 8 + }, + "render-051": { + "answer": "iss-tern-ghosting-on-moving-saltglass--2yr3q6zf476wajqp", + "tokens": 11365, + "fixed_tokens": 1096, + "processed_tokens": 120471, + "tool_calls": 9, + "turns": 10 + }, + "render-052": { + "answer": "61 percent", + "tokens": 7570, + "fixed_tokens": 1100, + "processed_tokens": 58195, + "tool_calls": 5, + "turns": 6 + }, + "render-053": { + "answer": "every frame", + "tokens": 11056, + "fixed_tokens": 1098, + "processed_tokens": 94718, + "tool_calls": 9, + "turns": 8 + }, + "render-054": { + "answer": "debug-gpu-validation", + "tokens": 12513, + "fixed_tokens": 1094, + "processed_tokens": 128723, + "tool_calls": 11, + "turns": 10 + }, + "render-055": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-with-the-reac--81tcpzvhwzcy6wyn", + "tokens": 12168, + "fixed_tokens": 1100, + "processed_tokens": 122083, + "tool_calls": 9, + "turns": 10 + }, + "render-056": { + "answer": "208,000 meshlets", + "tokens": 12165, + "fixed_tokens": 1101, + "processed_tokens": 151100, + "tool_calls": 11, + "turns": 12 + }, + "render-057": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 13983, + "fixed_tokens": 1093, + "processed_tokens": 141751, + "tool_calls": 10, + "turns": 11 + }, + "render-058": { + "answer": "iss-which-upscaler-should-lantern-ship--g8c2y9tsr856tn2v", + "tokens": 13954, + "fixed_tokens": 1089, + "processed_tokens": 125829, + "tool_calls": 9, + "turns": 10 + }, + "render-059": { + "answer": "fnd-sdf-gi-costs-4-7-ms-on-meridian--sm19rt2f274ajsvr", + "tokens": 11318, + "fixed_tokens": 1096, + "processed_tokens": 94682, + "tool_calls": 7, + "turns": 8 + }, + "render-060": { + "answer": "0.31", + "tokens": 12222, + "fixed_tokens": 1096, + "processed_tokens": 151125, + "tool_calls": 11, + "turns": 12 + }, + "render-061": { + "answer": "con-ui-composites-at-a-fixed-203-nits-in-hdr--de6aft7cvhdezn5s", + "tokens": 8803, + "fixed_tokens": 1094, + "processed_tokens": 85465, + "tool_calls": 7, + "turns": 8 + }, + "render-062": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 9887, + "fixed_tokens": 1094, + "processed_tokens": 98064, + "tool_calls": 8, + "turns": 9 + }, + "render-063": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 7981, + "fixed_tokens": 1095, + "processed_tokens": 59585, + "tool_calls": 5, + "turns": 6 + }, + "render-064": { + "answer": "dec-use-a-cached-shadow-atlas-for-local-lights--8p6f2v774he38t5j", + "tokens": 10089, + "fixed_tokens": 1099, + "processed_tokens": 88524, + "tool_calls": 8, + "turns": 8 + }, + "render-065": { + "answer": "dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e", + "tokens": 9270, + "fixed_tokens": 1091, + "processed_tokens": 99365, + "tool_calls": 8, + "turns": 9 + }, + "render-066": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 9202, + "fixed_tokens": 1094, + "processed_tokens": 72856, + "tool_calls": 6, + "turns": 7 + }, + "render-067": { + "answer": 4, + "tokens": 19957, + "fixed_tokens": 1096, + "processed_tokens": 141174, + "tool_calls": 8, + "turns": 9 + }, + "render-068": { + "answer": [ + "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "dec-lock-caustic-photon-jitter-to-the-tide-phase--85x10k4h11t8nx7c", + "dec-use-the-probe-grid-alone-in-120-hz-performance-m--6dzm1nt40pjfpyae", + "dec-trace-sdf-gi-at-quarter-resolution-in-performanc--06cqt3mqn3fg5h6e", + "dec-read-paper-white-from-the-meridian-system-settin--7crbx73gqw1hn7xq", + "dec-add-an-in-game-paper-white-slider-from-80-to-400--1phfvxby983z8tjf" + ], + "tokens": 14912, + "fixed_tokens": 1099, + "processed_tokens": 103132, + "tool_calls": 9, + "turns": 8 + }, + "render-069": { + "answer": "iss-shadow-pages-run-out-at-spring-low-tide--f4fgy6b44jbkrsmh", + "tokens": 29846, + "fixed_tokens": 1094, + "processed_tokens": 234206, + "tool_calls": 11, + "turns": 12 + }, + "render-070": { + "answer": [ + "dec-use-signed-distance-field-gi--yn366zmp3k7snd9w", + "dec-enable-meshlet-cone-culling-on-all-platforms--cpyweprmzcqhqsv2" + ], + "tokens": 80503, + "fixed_tokens": 1090, + "processed_tokens": 860982, + "tool_calls": 26, + "turns": 23 + }, + "render-071": { + "answer": 3, + "tokens": 13420, + "fixed_tokens": 1096, + "processed_tokens": 118700, + "tool_calls": 9, + "turns": 10 + }, + "render-072": { + "answer": [ + "rsk-sdf-volumes-outgrow-the-render-target-memory-bud--xfd2pfztvskdcd2c", + "rsk-artists-cannot-match-concept-paintings-under-the--6cfh9vkamn0kbtcx" + ], + "tokens": 9445, + "fixed_tokens": 1093, + "processed_tokens": 56952, + "tool_calls": 5, + "turns": 6 + }, + "render-073": { + "answer": [ + "rsk-vsm-page-pool-overflows-when-tides-expose-the-fl--x1r161nrdngra4ty", + "rsk-gi-cannot-fit-the-120-hz-performance-mode--tqt3fj839dqfq5s7", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 14466, + "fixed_tokens": 1094, + "processed_tokens": 110647, + "tool_calls": 12, + "turns": 9 + }, + "render-074": { + "answer": 5, + "tokens": 7538, + "fixed_tokens": 1093, + "processed_tokens": 53574, + "tool_calls": 5, + "turns": 6 + }, + "render-075": { + "answer": 3, + "tokens": 26485, + "fixed_tokens": 1097, + "processed_tokens": 445996, + "tool_calls": 23, + "turns": 24 + }, + "render-076": { + "answer": false, + "tokens": 14811, + "fixed_tokens": 1092, + "processed_tokens": 124352, + "tool_calls": 9, + "turns": 10 + }, + "render-077": { + "answer": true, + "tokens": 22330, + "fixed_tokens": 1097, + "processed_tokens": 201840, + "tool_calls": 12, + "turns": 13 + }, + "render-078": { + "answer": [ + "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 20659, + "fixed_tokens": 1099, + "processed_tokens": 223724, + "tool_calls": 26, + "turns": 14 + }, + "render-079": { + "answer": 6, + "tokens": 80197, + "fixed_tokens": 1092, + "processed_tokens": 1086247, + "tool_calls": 25, + "turns": 26 + }, + "render-080": { + "answer": "6 versions", + "tokens": 11236, + "fixed_tokens": 1095, + "processed_tokens": 91379, + "tool_calls": 7, + "turns": 8 + }, + "render-081": { + "answer": "unknown", + "tokens": 15285, + "fixed_tokens": 1090, + "processed_tokens": 106426, + "tool_calls": 9, + "turns": 8 + }, + "render-082": { + "answer": false, + "tokens": 12948, + "fixed_tokens": 1095, + "processed_tokens": 69476, + "tool_calls": 9, + "turns": 6 + }, + "render-083": { + "answer": 5, + "tokens": 78323, + "fixed_tokens": 1092, + "processed_tokens": 919974, + "tool_calls": 24, + "turns": 23 + }, + "tools-001": { + "answer": "completed", + "tokens": 8873, + "fixed_tokens": 1094, + "processed_tokens": 61223, + "tool_calls": 5, + "turns": 6 + }, + "tools-002": { + "answer": "47 minutes", + "tokens": 10616, + "fixed_tokens": 1094, + "processed_tokens": 76182, + "tool_calls": 6, + "turns": 7 + }, + "tools-003": { + "answer": "6 min 40 s", + "tokens": 8049, + "fixed_tokens": 1099, + "processed_tokens": 59660, + "tool_calls": 5, + "turns": 6 + }, + "tools-004": { + "answer": "unknown", + "tokens": 21970, + "fixed_tokens": 1098, + "processed_tokens": 163314, + "tool_calls": 9, + "turns": 10 + }, + "tools-005": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 10164, + "fixed_tokens": 1095, + "processed_tokens": 101217, + "tool_calls": 8, + "turns": 9 + }, + "tools-006": { + "answer": "dec-build-a-content-addressed-shader-compile-cache-s--e65rq9439k7xz5zh", + "tokens": 9224, + "fixed_tokens": 1092, + "processed_tokens": 74120, + "tool_calls": 7, + "turns": 7 + }, + "tools-007": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 29605, + "fixed_tokens": 1098, + "processed_tokens": 235467, + "tool_calls": 11, + "turns": 12 + }, + "tools-008": { + "answer": "dec-key-the-shader-cache-on-raw-source-text-and-comp--hp6awmx3gq5rn9nc", + "tokens": 9964, + "fixed_tokens": 1098, + "processed_tokens": 87551, + "tool_calls": 7, + "turns": 8 + }, + "tools-009": { + "answer": "92.4%", + "tokens": 7551, + "fixed_tokens": 1094, + "processed_tokens": 57843, + "tool_calls": 5, + "turns": 6 + }, + "tools-010": { + "answer": "dec-cap-each-farm-agent-at-6-parallel-shader-compile--q5w4sfjm73tsj8x8", + "tokens": 12820, + "fixed_tokens": 1101, + "processed_tokens": 109531, + "tool_calls": 10, + "turns": 9 + }, + "tools-011": { + "answer": "fnd-two-compiler-builds-with-the-same-version-string--gmgr4ha0ndxqmcn9", + "tokens": 12080, + "fixed_tokens": 1099, + "processed_tokens": 149902, + "tool_calls": 11, + "turns": 12 + }, + "tools-012": { + "answer": 3, + "tokens": 8151, + "fixed_tokens": 1096, + "processed_tokens": 83199, + "tool_calls": 8, + "turns": 9 + }, + "tools-013": { + "answer": 8, + "tokens": 73572, + "fixed_tokens": 1095, + "processed_tokens": 1654527, + "tool_calls": 38, + "turns": 39 + }, + "tools-014": { + "answer": false, + "tokens": 13318, + "fixed_tokens": 1095, + "processed_tokens": 123114, + "tool_calls": 9, + "turns": 10 + }, + "tools-015": { + "answer": [ + "iss-whether-the-audio-bank-packer-should-load-on-a-w--sqzp2rwp6s01w3b0", + "iss-whether-the-level-editor-autosave-needs-its-own--rkvmnbqe7w644zkr", + "iss-whether-the-artist-sandbox-should-load-on-a-work--73j819anqct8etfn", + "iss-whether-the-review-bot-should-load-on-a-worker-t--rtsjwvsksjdq15v1", + "iss-whether-the-level-editor-autosave-should-be-cach--radcjs4hyvh7p55q", + "iss-whether-the-crash-symbol-server-should-load-on-a--qvyphhgnsnm2cvqb", + "iss-whether-the-license-scanner-should-be-cached-bet--nqhd89czbfat9rya", + "iss-whether-the-doc-site-should-load-on-a-worker-thr--qxswe0pg52ekky0z", + "iss-whether-the-asset-search-should-load-on-a-worker--f9gpv3z522x6p6h4", + "iss-whether-the-changelog-bot-should-load-on-a-worke--jgtdn2f02p8fx5nr", + "iss-whether-the-review-bot-needs-its-own-memory-pool--pg3m8g9gn3y6v3kq", + "iss-whether-the-changelog-bot-should-skip-work-when--p2vrp5d682p3jrc3", + "iss-whether-the-patch-differ-should-load-on-a-worker--cysv4n56q1qn7t2k", + "iss-whether-the-artist-sandbox-should-skip-work-when--11tqzws72gms2h8x", + "iss-whether-the-audio-bank-packer-needs-its-own-memo--qjrqrseast8d5mhf", + "iss-whether-the-changelog-bot-needs-its-own-memory-p--mw8zejz6hk5m60x0", + "iss-whether-the-crash-symbol-server-should-ship-in-t--bhq91ajavah18wsk", + "iss-whether-the-asset-search-should-load-on-a-worker--80b27kqh36rpafdt", + "iss-whether-the-doc-site-should-ship-in-the-first-pa--wcdkq5zab7dxe5yj", + "iss-whether-the-patch-differ-should-be-cached-betwee--28b5zqb1ketwzf0r", + "iss-whether-the-test-runner-should-load-on-a-worker--aygay1r0qj4mzf78", + "iss-whether-the-texture-viewer-should-be-split-into--1jhgyqb70h5x7m7b", + "iss-whether-the-level-editor-autosave-should-skip-wo--fyysrcv6fvbsvcsy", + "iss-whether-the-level-editor-autosave-should-be-spli--ncnwh0csw671fs0q", + "iss-whether-the-asset-search-should-ship-in-the-firs--hsb2avhx613prxew", + "iss-whether-the-test-runner-needs-its-own-memory-poo--bs6arq0qv37d6wmx", + "iss-whether-the-artist-sandbox-should-be-split-into--ek47ca2vgmrjanp8" + ], + "tokens": 47583, + "fixed_tokens": 1094, + "processed_tokens": 342354, + "tool_calls": 12, + "turns": 13 + }, + "tools-016": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 17548, + "fixed_tokens": 1100, + "processed_tokens": 127275, + "tool_calls": 8, + "turns": 9 + }, + "world-001": { + "answer": "64 KB", + "tokens": 7689, + "fixed_tokens": 1092, + "processed_tokens": 58581, + "tool_calls": 5, + "turns": 6 + }, + "world-002": { + "answer": "5.1 ms", + "tokens": 8022, + "fixed_tokens": 1093, + "processed_tokens": 59462, + "tool_calls": 5, + "turns": 6 + }, + "world-003": { + "answer": "1 crash per 900 script waits", + "tokens": 8096, + "fixed_tokens": 1093, + "processed_tokens": 59421, + "tool_calls": 5, + "turns": 6 + }, + "world-004": { + "answer": "4,100 file opens", + "tokens": 8017, + "fixed_tokens": 1095, + "processed_tokens": 59441, + "tool_calls": 5, + "turns": 6 + }, + "world-005": { + "answer": "11 m", + "tokens": 7860, + "fixed_tokens": 1092, + "processed_tokens": 59000, + "tool_calls": 5, + "turns": 6 + }, + "world-006": { + "answer": "3.4 reversals per minute", + "tokens": 9567, + "fixed_tokens": 1093, + "processed_tokens": 88123, + "tool_calls": 7, + "turns": 8 + }, + "world-007": { + "answer": "260 MB", + "tokens": 9799, + "fixed_tokens": 1090, + "processed_tokens": 87568, + "tool_calls": 7, + "turns": 8 + }, + "world-008": { + "answer": "10 Hz", + "tokens": 11487, + "fixed_tokens": 1097, + "processed_tokens": 78615, + "tool_calls": 6, + "turns": 7 + }, + "world-009": { + "answer": "0.3 ms", + "tokens": 7561, + "fixed_tokens": 1089, + "processed_tokens": 58236, + "tool_calls": 5, + "turns": 6 + }, + "world-010": { + "answer": "29 GB/s", + "tokens": 9638, + "fixed_tokens": 1092, + "processed_tokens": 75704, + "tool_calls": 6, + "turns": 7 + }, + "world-011": { + "answer": "fnd-shipping-entity-names-costs-57-mb-at-full-popula--6w7gjvy7wdmdx9qx", + "tokens": 9253, + "fixed_tokens": 1079, + "processed_tokens": 85667, + "tool_calls": 7, + "turns": 8 + }, + "world-012": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 7576, + "fixed_tokens": 1097, + "processed_tokens": 58276, + "tool_calls": 5, + "turns": 6 + }, + "world-013": { + "answer": "unknown", + "tokens": 13552, + "fixed_tokens": 1093, + "processed_tokens": 82268, + "tool_calls": 6, + "turns": 7 + }, + "world-014": { + "answer": "unknown", + "tokens": 35425, + "fixed_tokens": 1092, + "processed_tokens": 268073, + "tool_calls": 11, + "turns": 12 + }, + "world-015": { + "answer": "3.0 ms per frame", + "tokens": 7313, + "fixed_tokens": 1097, + "processed_tokens": 57729, + "tool_calls": 5, + "turns": 6 + }, + "world-016": { + "answer": "unknown", + "tokens": 6895, + "fixed_tokens": 1096, + "processed_tokens": 64087, + "tool_calls": 6, + "turns": 7 + }, + "world-017": { + "answer": "dec-partition-the-world-into-128-m-cells--372b6czbjf29ymaw", + "tokens": 16678, + "fixed_tokens": 1092, + "processed_tokens": 103430, + "tool_calls": 8, + "turns": 7 + }, + "world-018": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 10726, + "fixed_tokens": 1097, + "processed_tokens": 88862, + "tool_calls": 7, + "turns": 8 + }, + "world-019": { + "answer": "fnd-tide-priority-reads-starve-player-proximity-load--by0d7b0bk39xz3rk", + "tokens": 7748, + "fixed_tokens": 1097, + "processed_tokens": 58704, + "tool_calls": 5, + "turns": 6 + }, + "world-020": { + "answer": "dec-schedule-streaming-reads-by-deadline-instead-of--q2x5aknabgcssdj2", + "tokens": 7488, + "fixed_tokens": 1096, + "processed_tokens": 57912, + "tool_calls": 5, + "turns": 6 + }, + "world-021": { + "answer": "dec-cut-navigation-tiles-to-64-m-so-three-fit-along--snvepk8ygeb3qsjs", + "tokens": 7608, + "fixed_tokens": 1089, + "processed_tokens": 58465, + "tool_calls": 5, + "turns": 6 + }, + "world-022": { + "answer": "fnd-41-of-alpha-crash-reports-have-no-usable-stack--35a50b07ferhf13m", + "tokens": 7557, + "fixed_tokens": 1092, + "processed_tokens": 58008, + "tool_calls": 5, + "turns": 6 + }, + "world-023": { + "answer": "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "tokens": 9856, + "fixed_tokens": 1095, + "processed_tokens": 88119, + "tool_calls": 7, + "turns": 8 + }, + "world-024": { + "answer": "dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew", + "tokens": 7683, + "fixed_tokens": 1097, + "processed_tokens": 58276, + "tool_calls": 5, + "turns": 6 + }, + "world-025": { + "answer": "con-streaming-pool-budget--kca88sgh2j6sxvd5", + "tokens": 9696, + "fixed_tokens": 1094, + "processed_tokens": 99969, + "tool_calls": 8, + "turns": 9 + }, + "world-026": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 7968, + "fixed_tokens": 1091, + "processed_tokens": 59521, + "tool_calls": 5, + "turns": 6 + }, + "world-027": { + "answer": "160 MB/s", + "tokens": 7725, + "fixed_tokens": 1096, + "processed_tokens": 58618, + "tool_calls": 5, + "turns": 6 + }, + "world-028": { + "answer": "fnd-streaming-tide-heights-costs-46-kb-s-per-peer--knxqtyqh6xcz1tp9", + "tokens": 7296, + "fixed_tokens": 1090, + "processed_tokens": 57317, + "tool_calls": 5, + "turns": 6 + }, + "world-029": { + "answer": "dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h", + "tokens": 9839, + "fixed_tokens": 1092, + "processed_tokens": 87015, + "tool_calls": 7, + "turns": 8 + }, + "world-030": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 11316, + "fixed_tokens": 1098, + "processed_tokens": 90187, + "tool_calls": 9, + "turns": 8 + }, + "world-031": { + "answer": "1.5 GB/s", + "tokens": 8634, + "fixed_tokens": 1090, + "processed_tokens": 61779, + "tool_calls": 5, + "turns": 6 + }, + "world-032": { + "answer": "fnd-float-tide-heights-drift-apart-between-pc-and-me--7jb0fg4b5xzr7q5e", + "tokens": 9543, + "fixed_tokens": 1088, + "processed_tokens": 86233, + "tool_calls": 7, + "turns": 8 + }, + "world-033": { + "answer": "dec-save-player-changes-as-deltas-against-the-tide-f--n38scy1yaarsr8cw", + "tokens": 18138, + "fixed_tokens": 1093, + "processed_tokens": 165365, + "tool_calls": 12, + "turns": 11 + }, + "world-034": { + "answer": "fnd-cross-platform-cloud-saves-need-platform-agreeme--4qgfc1a4nq4padvz", + "tokens": 15793, + "fixed_tokens": 1092, + "processed_tokens": 127594, + "tool_calls": 9, + "turns": 10 + }, + "world-035": { + "answer": false, + "tokens": 8166, + "fixed_tokens": 1091, + "processed_tokens": 58972, + "tool_calls": 5, + "turns": 6 + }, + "world-036": { + "answer": "Kestrel retired the fiber scheduler for engine jobs on 2026-04-14 (record dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h) because 41% of alpha crash reports had no usable stack, which blocked crash triage for beta, and a continuation-based thread pool kept workers 93% busy, close to the 95% fibers gave.", + "tokens": 14678, + "fixed_tokens": 1098, + "processed_tokens": 94385, + "tool_calls": 6, + "turns": 7 + }, + "world-037": { + "answer": "fnd-hot-reload-drops-suspended-gull-coroutines--va0fn0svryzwbzws", + "tokens": 9243, + "fixed_tokens": 1097, + "processed_tokens": 62178, + "tool_calls": 7, + "turns": 6 + }, + "world-038": { + "answer": "3", + "tokens": 9727, + "fixed_tokens": 1087, + "processed_tokens": 87257, + "tool_calls": 7, + "turns": 8 + }, + "world-039": { + "answer": "1.2 GB/s", + "tokens": 8439, + "fixed_tokens": 1093, + "processed_tokens": 61006, + "tool_calls": 5, + "turns": 6 + }, + "world-040": { + "answer": "unknown", + "tokens": 8157, + "fixed_tokens": 1098, + "processed_tokens": 60506, + "tool_calls": 5, + "turns": 6 + }, + "world-041": { + "answer": "61%", + "tokens": 9241, + "fixed_tokens": 1097, + "processed_tokens": 72343, + "tool_calls": 6, + "turns": 7 + }, + "world-042": { + "answer": "5.4 s", + "tokens": 7525, + "fixed_tokens": 1091, + "processed_tokens": 58179, + "tool_calls": 5, + "turns": 6 + }, + "world-043": { + "answer": "5.8 min", + "tokens": 7808, + "fixed_tokens": 1091, + "processed_tokens": 58823, + "tool_calls": 5, + "turns": 6 + }, + "world-044": { + "answer": "dec-move-engine-jobs-to-fibers--qjkgnm8cex8ydzfq", + "tokens": 14634, + "fixed_tokens": 1087, + "processed_tokens": 219992, + "tool_calls": 15, + "turns": 16 + }, + "world-045": { + "answer": "dec-run-engine-jobs-on-fibers-and-pin-gull-scripts-t--rw28cbqfd07rd9te", + "tokens": 35992, + "fixed_tokens": 1091, + "processed_tokens": 315064, + "tool_calls": 14, + "turns": 15 + }, + "world-046": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 15163, + "fixed_tokens": 1091, + "processed_tokens": 128203, + "tool_calls": 10, + "turns": 9 + }, + "world-047": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 15530, + "fixed_tokens": 1090, + "processed_tokens": 136751, + "tool_calls": 10, + "turns": 11 + }, + "world-048": { + "answer": "dec-compile-gull-to-bytecode-for-a-register-vm--jrqje646fpevp5f3", + "tokens": 7863, + "fixed_tokens": 1092, + "processed_tokens": 59332, + "tool_calls": 5, + "turns": 6 + }, + "world-049": { + "answer": "22%", + "tokens": 7946, + "fixed_tokens": 1096, + "processed_tokens": 59500, + "tool_calls": 5, + "turns": 6 + }, + "world-050": { + "answer": "fnd-alpha-content-needs-6-2-ms-of-gull-time-per-fram--4mekgdpsdv5nk085", + "tokens": 14264, + "fixed_tokens": 1092, + "processed_tokens": 144552, + "tool_calls": 10, + "turns": 11 + }, + "world-051": { + "answer": "Because integer adds and multiplies in 32.32 fixed point give identical bits on every CPU and compiler, while float tides drifted apart between machines within 6 minutes; the strict IEEE float decision was rejected in favor of this (dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e).", + "tokens": 9704, + "fixed_tokens": 1094, + "processed_tokens": 62146, + "tool_calls": 7, + "turns": 6 + }, + "world-052": { + "answer": "Dara Whitlock", + "tokens": 7973, + "fixed_tokens": 1093, + "processed_tokens": 59479, + "tool_calls": 5, + "turns": 6 + }, + "world-053": { + "answer": "1.9 ms", + "tokens": 9584, + "fixed_tokens": 1094, + "processed_tokens": 86388, + "tool_calls": 7, + "turns": 8 + }, + "world-054": { + "answer": "Marta Lindqvist", + "tokens": 12013, + "fixed_tokens": 1092, + "processed_tokens": 149911, + "tool_calls": 11, + "turns": 12 + }, + "world-055": { + "answer": "Kelp stays because the cold-cache rerun that favored Zstandard tested the 4 MB dictionary; shrinking Kelp's dictionary to 1 MB makes it beat Zstandard cold on Skiff by 1.7x, per dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew.", + "tokens": 8664, + "fixed_tokens": 1092, + "processed_tokens": 61065, + "tool_calls": 5, + "turns": 6 + }, + "world-056": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 13839, + "fixed_tokens": 1100, + "processed_tokens": 114790, + "tool_calls": 10, + "turns": 9 + }, + "world-057": { + "answer": "tile [212, 87]", + "tokens": 12489, + "fixed_tokens": 1096, + "processed_tokens": 180246, + "tool_calls": 13, + "turns": 14 + }, + "world-058": { + "answer": "1.1 GB/s", + "tokens": 12150, + "fixed_tokens": 1096, + "processed_tokens": 99098, + "tool_calls": 9, + "turns": 8 + }, + "world-059": { + "answer": "dec-derive-the-missing-harmonic-phase-from-each-save--gbqrvvgyq2gy2gnw", + "tokens": 12475, + "fixed_tokens": 1093, + "processed_tokens": 126670, + "tool_calls": 11, + "turns": 10 + }, + "world-060": { + "answer": "dec-partition-the-world-into-192-m-cells--bk9r1c5jwb6fh8tc", + "tokens": 9402, + "fixed_tokens": 1095, + "processed_tokens": 86585, + "tool_calls": 7, + "turns": 8 + }, + "world-061": { + "answer": "236 MB", + "tokens": 13968, + "fixed_tokens": 1095, + "processed_tokens": 189445, + "tool_calls": 13, + "turns": 14 + }, + "world-062": { + "answer": "310 MB", + "tokens": 14449, + "fixed_tokens": 1091, + "processed_tokens": 156928, + "tool_calls": 11, + "turns": 12 + }, + "world-063": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 23044, + "fixed_tokens": 1091, + "processed_tokens": 248058, + "tool_calls": 14, + "turns": 15 + }, + "world-064": { + "answer": [ + "dec-force-strict-ieee-float-mode-for-the-tide-simula--ecwmqt7ncjngysqw", + "dec-have-the-host-stream-tide-heights-to-every-peer--a3rb1seedn2bj7k8" + ], + "tokens": 16536, + "fixed_tokens": 1097, + "processed_tokens": 164978, + "tool_calls": 12, + "turns": 12 + }, + "world-065": { + "answer": "3.1 ms", + "tokens": 14464, + "fixed_tokens": 1093, + "processed_tokens": 192394, + "tool_calls": 13, + "turns": 14 + }, + "world-066": { + "answer": "46 reads", + "tokens": 28058, + "fixed_tokens": 1089, + "processed_tokens": 438089, + "tool_calls": 19, + "turns": 20 + }, + "world-067": { + "answer": "1.2 GB/s", + "tokens": 10427, + "fixed_tokens": 1096, + "processed_tokens": 91685, + "tool_calls": 7, + "turns": 8 + }, + "world-068": { + "answer": true, + "tokens": 15983, + "fixed_tokens": 1091, + "processed_tokens": 192459, + "tool_calls": 14, + "turns": 15 + }, + "world-069": { + "answer": 6, + "tokens": 82850, + "fixed_tokens": 1094, + "processed_tokens": 835574, + "tool_calls": 20, + "turns": 21 + }, + "world-070": { + "answer": 3, + "tokens": 10758, + "fixed_tokens": 1091, + "processed_tokens": 102291, + "tool_calls": 8, + "turns": 9 + }, + "world-071": { + "answer": [ + "dec-hot-reload-gull-functions-by-swapping-bytecode-a--8dxnkg6h8vmcsr42", + "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "dec-evict-cells-by-distance-and-tide-forecast--fyp1jp366pjjzxcg" + ], + "tokens": 64299, + "fixed_tokens": 1094, + "processed_tokens": 901109, + "tool_calls": 24, + "turns": 25 + }, + "world-072": { + "answer": [ + "dec-compress-streamed-assets-with-lz4--v1pfsgfp0b3yjv4s", + "dec-compress-streamed-assets-with-zstandard--bs1t15cygc51cpje" + ], + "tokens": 31228, + "fixed_tokens": 1089, + "processed_tokens": 246360, + "tool_calls": 12, + "turns": 13 + }, + "world-073": { + "answer": false, + "tokens": 9204, + "fixed_tokens": 1100, + "processed_tokens": 60923, + "tool_calls": 7, + "turns": 6 + }, + "world-074": { + "answer": 7, + "tokens": 9375, + "fixed_tokens": 1098, + "processed_tokens": 70630, + "tool_calls": 6, + "turns": 7 + }, + "world-075": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--8679sjn071ycsbpv", + "dec-partition-the-world-into-192-meter-cells--q62ddwdv2n8wxay6", + "dec-return-engine-jobs-to-the-thread-pool-with-conti--547qbweqvrp1j5g3" + ], + "tokens": 6362, + "fixed_tokens": 1102, + "processed_tokens": 51100, + "tool_calls": 5, + "turns": 6 + }, + "world-076": { + "answer": "unknown", + "tokens": 32852, + "fixed_tokens": 1094, + "processed_tokens": 639923, + "tool_calls": 29, + "turns": 27 + }, + "world-077": { + "answer": 251, + "tokens": 29557, + "fixed_tokens": 1097, + "processed_tokens": 466625, + "tool_calls": 19, + "turns": 20 + }, + "world-078": { + "answer": [ + "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "rsk-reworking-the-read-scheduler-may-slip-the-coast--9jzvs7etvb5jxqtp" + ], + "tokens": 10823, + "fixed_tokens": 1100, + "processed_tokens": 59208, + "tool_calls": 5, + "turns": 6 + }, + "world-079": { + "answer": [ + "con-no-runtime-code-generation-on-consoles--6cf25zqmsvqkdd73", + "con-tide-state-must-be-bit-identical-on-every-peer--zrpnp8qjyxzkas7t", + "con-saves-from-beta-onward-must-load-in-every-later--v6z856sxtgc5sb09", + "con-streaming-pool-budget--8czwb4fjr5p5vpy7" + ], + "tokens": 10777, + "fixed_tokens": 1092, + "processed_tokens": 67506, + "tool_calls": 6, + "turns": 7 + }, + "world-080": { + "answer": [ + "iss-cross-platform-cloud-saves-are-unplanned--5bk0k4wrgj42ztfy", + "iss-gull-hot-reload-drops-suspended-coroutines--z7j2gmyny6gdxd07", + "iss-gull-hot-reload-rejects-struct-layout-changes--3v2w2f9hg2gwy6aj" + ], + "tokens": 11182, + "fixed_tokens": 1096, + "processed_tokens": 96929, + "tool_calls": 12, + "turns": 9 + }, + "world-081": { + "answer": "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "tokens": 37257, + "fixed_tokens": 1099, + "processed_tokens": 651569, + "tool_calls": 34, + "turns": 26 + } + } +} diff --git a/benchmarks/proof-recall/runs/x10-find-r2-nobrief.json b/benchmarks/proof-recall/runs/x10-find-r2-nobrief.json new file mode 100644 index 00000000..dfbfee7a --- /dev/null +++ b/benchmarks/proof-recall/runs/x10-find-r2-nobrief.json @@ -0,0 +1,2255 @@ +{ + "run": { + "name": "x10-find-r2-nobrief", + "protocol": "one fresh session per question", + "models": ["claude-haiku-5-5"], + "find_hash": "e5be653299abe9b2329822d98ffe09e90ee7cfe96f8711c813a51dc6500b887f" + }, + "answers": { + "ambiguous-001": { + "answer": "unknown", + "tokens": 53227, + "fixed_tokens": 1116, + "processed_tokens": 636896, + "tool_calls": 21, + "turns": 18 + }, + "ambiguous-002": { + "answer": "unknown", + "tokens": 21055, + "fixed_tokens": 1114, + "processed_tokens": 205186, + "tool_calls": 11, + "turns": 12 + }, + "ambiguous-003": { + "answer": "92.4%", + "tokens": 41137, + "fixed_tokens": 1119, + "processed_tokens": 1078318, + "tool_calls": 35, + "turns": 35 + }, + "ambiguous-004": { + "answer": "unknown", + "tokens": 45669, + "fixed_tokens": 1115, + "processed_tokens": 734305, + "tool_calls": 23, + "turns": 24 + }, + "ambiguous-005": { + "answer": "6 weeks", + "tokens": 46508, + "fixed_tokens": 1115, + "processed_tokens": 719221, + "tool_calls": 25, + "turns": 24 + }, + "ambiguous-006": { + "answer": "unknown", + "tokens": 62221, + "fixed_tokens": 1112, + "processed_tokens": 2189767, + "tool_calls": 50, + "turns": 51 + }, + "ambiguous-007": { + "answer": "32 steps", + "tokens": 31735, + "fixed_tokens": 1115, + "processed_tokens": 599946, + "tool_calls": 23, + "turns": 24 + }, + "ambiguous-008": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 22249, + "fixed_tokens": 1112, + "processed_tokens": 184849, + "tool_calls": 10, + "turns": 11 + }, + "ambiguous-009": { + "answer": "3.0 ms", + "tokens": 41784, + "fixed_tokens": 1109, + "processed_tokens": 490981, + "tool_calls": 16, + "turns": 17 + }, + "ambiguous-010": { + "answer": "unknown", + "tokens": 49798, + "fixed_tokens": 1118, + "processed_tokens": 990898, + "tool_calls": 29, + "turns": 30 + }, + "ambiguous-011": { + "answer": "260 MB", + "tokens": 34320, + "fixed_tokens": 1110, + "processed_tokens": 321470, + "tool_calls": 13, + "turns": 14 + }, + "ambiguous-012": { + "answer": "11 m", + "tokens": 27245, + "fixed_tokens": 1112, + "processed_tokens": 265236, + "tool_calls": 12, + "turns": 13 + }, + "ambiguous-013": { + "answer": "61%", + "tokens": 41560, + "fixed_tokens": 1114, + "processed_tokens": 543696, + "tool_calls": 17, + "turns": 18 + }, + "ambiguous-014": { + "answer": "textures", + "tokens": 41505, + "fixed_tokens": 1112, + "processed_tokens": 781590, + "tool_calls": 26, + "turns": 27 + }, + "ambiguous-015": { + "answer": "unknown", + "tokens": 62603, + "fixed_tokens": 1109, + "processed_tokens": 1075086, + "tool_calls": 24, + "turns": 25 + }, + "ambiguous-016": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 29380, + "fixed_tokens": 1110, + "processed_tokens": 549607, + "tool_calls": 22, + "turns": 23 + }, + "ambiguous-017": { + "answer": "unknown", + "tokens": 78449, + "fixed_tokens": 1112, + "processed_tokens": 2740858, + "tool_calls": 49, + "turns": 50 + }, + "ambiguous-018": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 105408, + "fixed_tokens": 1111, + "processed_tokens": 1922177, + "tool_calls": 33, + "turns": 27 + }, + "ambiguous-019": { + "answer": "7.4 ms", + "tokens": 25745, + "fixed_tokens": 1116, + "processed_tokens": 194660, + "tool_calls": 10, + "turns": 11 + }, + "ambiguous-020": { + "answer": "5912 pages", + "tokens": 24728, + "fixed_tokens": 1116, + "processed_tokens": 451869, + "tool_calls": 22, + "turns": 23 + }, + "ambiguous-021": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 35918, + "fixed_tokens": 1114, + "processed_tokens": 641013, + "tool_calls": 23, + "turns": 24 + }, + "ambiguous-022": { + "answer": "unknown", + "tokens": 45929, + "fixed_tokens": 1114, + "processed_tokens": 410873, + "tool_calls": 16, + "turns": 15 + }, + "ambiguous-023": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 8875, + "fixed_tokens": 1119, + "processed_tokens": 67684, + "tool_calls": 6, + "turns": 7 + }, + "ambiguous-024": { + "answer": "unknown", + "tokens": 65479, + "fixed_tokens": 1110, + "processed_tokens": 999873, + "tool_calls": 27, + "turns": 22 + }, + "cross-001": { + "answer": "con-frame-budget--br0qszfg8fys908n", + "tokens": 22303, + "fixed_tokens": 1111, + "processed_tokens": 194832, + "tool_calls": 10, + "turns": 11 + }, + "cross-002": { + "answer": 2, + "tokens": 28010, + "fixed_tokens": 1108, + "processed_tokens": 119935, + "tool_calls": 19, + "turns": 8 + }, + "cross-003": { + "answer": 2, + "tokens": 21071, + "fixed_tokens": 1104, + "processed_tokens": 94848, + "tool_calls": 14, + "turns": 7 + }, + "cross-004": { + "answer": "dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e", + "tokens": 35832, + "fixed_tokens": 1110, + "processed_tokens": 383342, + "tool_calls": 14, + "turns": 15 + }, + "cross-005": { + "answer": "Rhea Valdés", + "tokens": 30611, + "fixed_tokens": 1110, + "processed_tokens": 565128, + "tool_calls": 25, + "turns": 26 + }, + "cross-006": { + "answer": "~19% (4.8 ms of the 25 ms Skiff 40 Hz frame)", + "tokens": 33185, + "fixed_tokens": 1102, + "processed_tokens": 581035, + "tool_calls": 22, + "turns": 23 + }, + "cross-007": { + "answer": "8 GB", + "tokens": 36623, + "fixed_tokens": 1102, + "processed_tokens": 456266, + "tool_calls": 17, + "turns": 18 + }, + "cross-008": { + "answer": "unknown", + "tokens": 42483, + "fixed_tokens": 1106, + "processed_tokens": 560151, + "tool_calls": 18, + "turns": 18 + }, + "cross-009": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 52263, + "fixed_tokens": 1106, + "processed_tokens": 1747307, + "tool_calls": 47, + "turns": 46 + }, + "cross-010": { + "answer": "unknown", + "tokens": 47272, + "fixed_tokens": 1108, + "processed_tokens": 804788, + "tool_calls": 26, + "turns": 22 + }, + "cross-011": { + "answer": false, + "tokens": 66247, + "fixed_tokens": 1104, + "processed_tokens": 894854, + "tool_calls": 20, + "turns": 21 + }, + "cross-012": { + "answer": "unknown", + "tokens": 53546, + "fixed_tokens": 1108, + "processed_tokens": 820321, + "tool_calls": 20, + "turns": 21 + }, + "cross-013": { + "answer": "Skiff runs Tern at the Performance preset, which saves 1.8 ms over Quality (2.1 ms vs 3.9 ms), per dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3 and fnd-tern-costs-3-9-ms-at-quality-and-2-1-ms-at-perfo--dkbskbqfgfwtryq0. I found no record tying the Meridian 120 Hz mode to Tern's lowest preset, so that part of the premise is unverified.", + "tokens": 40775, + "fixed_tokens": 1108, + "processed_tokens": 678590, + "tool_calls": 23, + "turns": 24 + }, + "cross-014": { + "answer": 3, + "tokens": 5467, + "fixed_tokens": 1105, + "processed_tokens": 40195, + "tool_calls": 4, + "turns": 5 + }, + "cross-015": { + "answer": "Two Efforts are no longer active. Shader compile cache and build farm is completed: dec-turn-the-shader-compile-cache-on-for-every-farm--4rr3b12b4sasdmaj (2026-03-20) turned the cache on for all nightly, on-demand and release farm builds, which is the end condition the Effort set. Co-op rollback netcode prototype is abandoned: dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b (2026-04-01) stopped rollback because re-running tides on every rolled-back frame costs 4.8 ms on Skiff.", + "tokens": 34701, + "fixed_tokens": 1113, + "processed_tokens": 190570, + "tool_calls": 12, + "turns": 9 + }, + "cross-016": { + "answer": "16.7 ms", + "tokens": 26989, + "fixed_tokens": 1108, + "processed_tokens": 181827, + "tool_calls": 9, + "turns": 10 + }, + "cross-017": { + "answer": false, + "tokens": 36180, + "fixed_tokens": 1101, + "processed_tokens": 361089, + "tool_calls": 15, + "turns": 15 + }, + "netcode-001": { + "answer": "abandoned", + "tokens": 18022, + "fixed_tokens": 1104, + "processed_tokens": 85954, + "tool_calls": 6, + "turns": 7 + }, + "netcode-002": { + "answer": "netcode/rollback-archive", + "tokens": 20161, + "fixed_tokens": 1105, + "processed_tokens": 173636, + "tool_calls": 10, + "turns": 11 + }, + "netcode-003": { + "answer": "unknown", + "tokens": 92518, + "fixed_tokens": 1107, + "processed_tokens": 2185155, + "tool_calls": 34, + "turns": 35 + }, + "netcode-004": { + "answer": "The team abandoned rollback because re-running the tide simulation on every rolled-back frame costs 4.8 ms on Skiff, about a fifth of the 25 ms frame, and the workarounds failed (shrinking the window, quarter-resolution tides, leaving tides out of the snapshot). The reason depends on the world Decision dec-run-the-tide-simulation-in-32-32-fixed-point--c5rjfcnmhne4h05e.", + "tokens": 25420, + "fixed_tokens": 1110, + "processed_tokens": 332194, + "tool_calls": 15, + "turns": 16 + }, + "netcode-005": { + "answer": "unknown", + "tokens": 26556, + "fixed_tokens": 1108, + "processed_tokens": 394575, + "tool_calls": 17, + "turns": 18 + }, + "netcode-006": { + "answer": "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "tokens": 9563, + "fixed_tokens": 1108, + "processed_tokens": 109920, + "tool_calls": 10, + "turns": 11 + }, + "netcode-007": { + "answer": "dec-abandon-rollback-netcode-for-co-op--46envrg0v4pdwv7b", + "tokens": 20831, + "fixed_tokens": 1109, + "processed_tokens": 137972, + "tool_calls": 8, + "turns": 9 + }, + "netcode-008": { + "answer": "dec-use-a-4-frame-rollback-window--pqe4hvbgmbp0dmh9", + "tokens": 30386, + "fixed_tokens": 1108, + "processed_tokens": 356593, + "tool_calls": 15, + "turns": 16 + }, + "netcode-009": { + "answer": "1.9 ms", + "tokens": 22464, + "fixed_tokens": 1109, + "processed_tokens": 268291, + "tool_calls": 13, + "turns": 14 + }, + "netcode-010": { + "answer": "dec-use-a-6-frame-rollback-window-with-1-frame-of-in--s53944zhkw103frk", + "tokens": 58663, + "fixed_tokens": 1109, + "processed_tokens": 1589277, + "tool_calls": 41, + "turns": 42 + }, + "netcode-011": { + "answer": "con-co-op-peers-must-agree-on-tide-state-bit-for-bit--4v0sdbc3t4ywzr3g", + "tokens": 69525, + "fixed_tokens": 1111, + "processed_tokens": 922654, + "tool_calls": 30, + "turns": 23 + }, + "netcode-012": { + "answer": "dec-re-simulate-tides-at-quarter-resolution-during-r--7ah25p0r0xt9719f", + "tokens": 27535, + "fixed_tokens": 1109, + "processed_tokens": 316693, + "tool_calls": 14, + "turns": 15 + }, + "netcode-013": { + "answer": [ + "dec-prototype-rollback-netcode-for-co-op--2tx2s2p3sfbwj8x4", + "fnd-rollback-re-simulates-a-4-frame-window-in-1-1-ms--xa30a1cez2brphyf", + "rsk-re-simulation-cost-could-blow-the-skiff-frame-bu--sk2njsp170a3qwyf" + ], + "tokens": 34204, + "fixed_tokens": 1112, + "processed_tokens": 284942, + "tool_calls": 26, + "turns": 13 + }, + "netcode-014": { + "answer": [ + "iss-late-join-during-a-tide-surge-desyncs-the-joinin--mya93va0n7s8g7cj" + ], + "tokens": 51170, + "fixed_tokens": 1108, + "processed_tokens": 415489, + "tool_calls": 15, + "turns": 14 + }, + "netcode-015": { + "answer": 2, + "tokens": 10647, + "fixed_tokens": 1105, + "processed_tokens": 91760, + "tool_calls": 8, + "turns": 9 + }, + "ports-001": { + "answer": "25 ms", + "tokens": 22834, + "fixed_tokens": 1111, + "processed_tokens": 223374, + "tool_calls": 12, + "turns": 12 + }, + "ports-002": { + "answer": "unknown", + "tokens": 50345, + "fixed_tokens": 1101, + "processed_tokens": 964903, + "tool_calls": 25, + "turns": 26 + }, + "ports-003": { + "answer": "61 MB/s", + "tokens": 72024, + "fixed_tokens": 1106, + "processed_tokens": 2362047, + "tool_calls": 53, + "turns": 54 + }, + "ports-004": { + "answer": "accepted", + "tokens": 18802, + "fixed_tokens": 1104, + "processed_tokens": 282932, + "tool_calls": 19, + "turns": 20 + }, + "ports-005": { + "answer": "unknown", + "tokens": 57335, + "fixed_tokens": 1106, + "processed_tokens": 1435551, + "tool_calls": 39, + "turns": 37 + }, + "ports-034": { + "answer": "unknown", + "tokens": 62515, + "fixed_tokens": 1107, + "processed_tokens": 1543580, + "tool_calls": 46, + "turns": 38 + }, + "ports-006": { + "answer": "con-skiff-memory-budget--nf7qbhsnmcn6v479", + "tokens": 18824, + "fixed_tokens": 1105, + "processed_tokens": 159378, + "tool_calls": 9, + "turns": 10 + }, + "ports-007": { + "answer": "SKF-UI-031", + "tokens": 19257, + "fixed_tokens": 1101, + "processed_tokens": 161440, + "tool_calls": 9, + "turns": 10 + }, + "ports-008": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 8260, + "fixed_tokens": 1103, + "processed_tokens": 86648, + "tool_calls": 8, + "turns": 9 + }, + "ports-009": { + "answer": "dec-load-skiff-fast-travel-without-a-loading-screen--v4qft5znysjv5f24", + "tokens": 34738, + "fixed_tokens": 1100, + "processed_tokens": 770086, + "tool_calls": 28, + "turns": 29 + }, + "ports-010": { + "answer": "con-frame-budget--wt5rfqh5z444g2cg", + "tokens": 51223, + "fixed_tokens": 1105, + "processed_tokens": 1350619, + "tool_calls": 38, + "turns": 39 + }, + "ports-011": { + "answer": "unknown", + "tokens": 84932, + "fixed_tokens": 1104, + "processed_tokens": 838566, + "tool_calls": 17, + "turns": 18 + }, + "ports-012": { + "answer": "Saltglass caustic patterns shimmered badly in motion at the 540p floor (visible in 7 of 10 dune views, versus 2 of 10 at 576p), so the team raised the floor to 576p (dec-raise-the-skiff-dynamic-resolution-floor-to-576p--fep9q7m08pp9f0bb, since superseded by a 504p decision).", + "tokens": 23278, + "fixed_tokens": 1104, + "processed_tokens": 272458, + "tool_calls": 13, + "turns": 14 + }, + "ports-013": { + "answer": "Ship Tern as Lantern's only upscaler", + "tokens": 41094, + "fixed_tokens": 1103, + "processed_tokens": 402534, + "tool_calls": 13, + "turns": 14 + }, + "ports-014": { + "answer": "dec-lower-the-skiff-dynamic-resolution-floor-to-504p--15esnt90gcn29enn", + "tokens": 65339, + "fixed_tokens": 1102, + "processed_tokens": 635503, + "tool_calls": 18, + "turns": 17 + }, + "ports-015": { + "answer": "textures", + "tokens": 18359, + "fixed_tokens": 1104, + "processed_tokens": 264161, + "tool_calls": 17, + "turns": 18 + }, + "ports-016": { + "answer": "dec-ship-prebuilt-pipeline-blobs-in-the-skiff-packag--3vjysjgy9qkvzebn", + "tokens": 34606, + "fixed_tokens": 1109, + "processed_tokens": 756732, + "tool_calls": 28, + "turns": 29 + }, + "ports-017": { + "answer": "fnd-skiff-microsd-slows-sharply-on-full-warm-cards--p1rgh36hat2egpz3", + "tokens": 46304, + "fixed_tokens": 1101, + "processed_tokens": 1051583, + "tool_calls": 28, + "turns": 29 + }, + "ports-018": { + "answer": "504p", + "tokens": 13243, + "fixed_tokens": 1108, + "processed_tokens": 129759, + "tool_calls": 10, + "turns": 11 + }, + "ports-019": { + "answer": "4 MB per minute", + "tokens": 43302, + "fixed_tokens": 1114, + "processed_tokens": 945911, + "tool_calls": 36, + "turns": 30 + }, + "ports-020": { + "answer": "fnd-skiff-sdk-1-2-takes-1-gb-back-from-games--a0ybdgjzw64hwfqa", + "tokens": 25789, + "fixed_tokens": 1109, + "processed_tokens": 341989, + "tool_calls": 16, + "turns": 17 + }, + "ports-021": { + "answer": "unknown", + "tokens": 75820, + "fixed_tokens": 1105, + "processed_tokens": 2645541, + "tool_calls": 56, + "turns": 57 + }, + "ports-022": { + "answer": "cit-skiff-sdk-1-2-release-notes--7gwwn8xxffrx6npe", + "tokens": 23130, + "fixed_tokens": 1102, + "processed_tokens": 269185, + "tool_calls": 13, + "turns": 14 + }, + "ports-023": { + "answer": "fnd-tern-costs-3-9-ms-at-quality-and-2-1-ms-at-perfo--dkbskbqfgfwtryq0", + "tokens": 27160, + "fixed_tokens": 1107, + "processed_tokens": 408130, + "tool_calls": 18, + "turns": 19 + }, + "ports-024": { + "answer": "fnd-skiff-firmware-1-2-rejects-pipeline-blobs-built--pernvgb2he7m3z3y", + "tokens": 33940, + "fixed_tokens": 1106, + "processed_tokens": 749209, + "tool_calls": 30, + "turns": 28 + }, + "ports-025": { + "answer": "dec-run-tern-on-skiff-at-the-performance-preset--74eqayfa83ybqpe3", + "tokens": 27736, + "fixed_tokens": 1108, + "processed_tokens": 273205, + "tool_calls": 12, + "turns": 13 + }, + "ports-026": { + "answer": "912 MB", + "tokens": 37191, + "fixed_tokens": 1101, + "processed_tokens": 918883, + "tool_calls": 32, + "turns": 33 + }, + "ports-027": { + "answer": [ + "rsk-tide-streaming-outruns-skiff-storage--6q8s3b383z2tyeyb", + "dec-stream-skiff-world-tiles-in-2-mb-reads--3jz9w9gvhhc7eppw" + ], + "tokens": 80471, + "fixed_tokens": 1110, + "processed_tokens": 1463994, + "tool_calls": 41, + "turns": 28 + }, + "ports-028": { + "answer": 2, + "tokens": 9311, + "fixed_tokens": 1104, + "processed_tokens": 67754, + "tool_calls": 6, + "turns": 7 + }, + "ports-029": { + "answer": 8, + "tokens": 45747, + "fixed_tokens": 1103, + "processed_tokens": 396759, + "tool_calls": 15, + "turns": 16 + }, + "ports-030": { + "answer": [ + "dec-cap-tide-storm-particles-at-6-000-on-skiff--resze52afjxyxys1", + "dec-drop-skiff-to-30-hz-while-a-tide-storm-is-on-scr--72vx1c8s4jrs6axc" + ], + "tokens": 9808, + "fixed_tokens": 1108, + "processed_tokens": 102369, + "tool_calls": 9, + "turns": 10 + }, + "ports-031": { + "answer": [ + "iss-meridian-cert-activity-cards-missing-for-two-bio--k214cmfra7pyvvzq", + "iss-skiff-cert-docked-30-hz-judder-on-some-tvs--9nq2xg9mzash474g" + ], + "tokens": 65005, + "fixed_tokens": 1105, + "processed_tokens": 1215390, + "tool_calls": 28, + "turns": 29 + }, + "ports-032": { + "answer": 2, + "tokens": 6572, + "fixed_tokens": 1105, + "processed_tokens": 61323, + "tool_calls": 6, + "turns": 7 + }, + "ports-033": { + "answer": "unknown", + "tokens": 32167, + "fixed_tokens": 1107, + "processed_tokens": 280770, + "tool_calls": 11, + "turns": 12 + }, + "render-001": { + "answer": "7.4 ms", + "tokens": 36530, + "fixed_tokens": 1099, + "processed_tokens": 222341, + "tool_calls": 11, + "turns": 10 + }, + "render-002": { + "answer": "3.4x", + "tokens": 22281, + "fixed_tokens": 1107, + "processed_tokens": 245187, + "tool_calls": 12, + "turns": 13 + }, + "render-003": { + "answer": "Tomasz Brandt", + "tokens": 16226, + "fixed_tokens": 1104, + "processed_tokens": 99487, + "tool_calls": 7, + "turns": 8 + }, + "render-004": { + "answer": "rejected", + "tokens": 34894, + "fixed_tokens": 1100, + "processed_tokens": 710041, + "tool_calls": 34, + "turns": 34 + }, + "render-005": { + "answer": true, + "tokens": 31599, + "fixed_tokens": 1098, + "processed_tokens": 396779, + "tool_calls": 16, + "turns": 17 + }, + "render-006": { + "answer": "unknown", + "tokens": 37719, + "fixed_tokens": 1102, + "processed_tokens": 425386, + "tool_calls": 17, + "turns": 18 + }, + "render-007": { + "answer": "2026-07-24", + "tokens": 29978, + "fixed_tokens": 1102, + "processed_tokens": 313685, + "tool_calls": 14, + "turns": 14 + }, + "render-008": { + "answer": "unknown", + "tokens": 36288, + "fixed_tokens": 1103, + "processed_tokens": 962250, + "tool_calls": 41, + "turns": 41 + }, + "render-009": { + "answer": "32 steps", + "tokens": 24267, + "fixed_tokens": 1100, + "processed_tokens": 231157, + "tool_calls": 12, + "turns": 13 + }, + "render-010": { + "answer": "unknown", + "tokens": 25643, + "fixed_tokens": 1102, + "processed_tokens": 194542, + "tool_calls": 10, + "turns": 11 + }, + "render-011": { + "answer": "fnd-sdf-gi-costs-2-4-ms-on-meridian--2w6c2pc5zmcsyjre", + "tokens": 14941, + "fixed_tokens": 1100, + "processed_tokens": 235165, + "tool_calls": 17, + "turns": 18 + }, + "render-012": { + "answer": "deferred", + "tokens": 53978, + "fixed_tokens": 1106, + "processed_tokens": 845052, + "tool_calls": 23, + "turns": 22 + }, + "render-013": { + "answer": "16-bit float per channel", + "tokens": 37541, + "fixed_tokens": 1103, + "processed_tokens": 528693, + "tool_calls": 19, + "turns": 20 + }, + "render-014": { + "answer": "52%", + "tokens": 27776, + "fixed_tokens": 1104, + "processed_tokens": 640685, + "tool_calls": 32, + "turns": 33 + }, + "render-015": { + "answer": "unknown", + "tokens": 63571, + "fixed_tokens": 1102, + "processed_tokens": 2108494, + "tool_calls": 59, + "turns": 60 + }, + "render-016": { + "answer": "1.2 km", + "tokens": 7601, + "fixed_tokens": 1102, + "processed_tokens": 83013, + "tool_calls": 8, + "turns": 9 + }, + "render-017": { + "answer": "unknown", + "tokens": 59486, + "fixed_tokens": 1103, + "processed_tokens": 662004, + "tool_calls": 18, + "turns": 19 + }, + "render-018": { + "answer": "unknown", + "tokens": 37229, + "fixed_tokens": 1103, + "processed_tokens": 403833, + "tool_calls": 14, + "turns": 15 + }, + "render-019": { + "answer": "dec-use-the-custom-brine-curve-for-tonemapping--p735g6wf2njare86", + "tokens": 29820, + "fixed_tokens": 1102, + "processed_tokens": 441393, + "tool_calls": 21, + "turns": 21 + }, + "render-020": { + "answer": "fnd-screen-space-refraction-loses-off-screen-content--k8zvmmdee7wg2fjy", + "tokens": 24324, + "fixed_tokens": 1099, + "processed_tokens": 604357, + "tool_calls": 34, + "turns": 35 + }, + "render-021": { + "answer": "dec-adopt-a-visibility-buffer-for-all-opaque-geometr--68vchzqgd1avw8q4", + "tokens": 20419, + "fixed_tokens": 1102, + "processed_tokens": 122585, + "tool_calls": 7, + "turns": 8 + }, + "render-022": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 9820, + "fixed_tokens": 1100, + "processed_tokens": 100876, + "tool_calls": 9, + "turns": 10 + }, + "render-023": { + "answer": "dec-capture-the-cliffs-and-dunes-benchmarks-on-every--3pqyr66yh4v6q676", + "tokens": 22016, + "fixed_tokens": 1096, + "processed_tokens": 213802, + "tool_calls": 11, + "turns": 12 + }, + "render-024": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 14868, + "fixed_tokens": 1097, + "processed_tokens": 115287, + "tool_calls": 11, + "turns": 9 + }, + "render-025": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-after-fp16-pa--p4gcmzw2a4bvzjev", + "tokens": 49181, + "fixed_tokens": 1101, + "processed_tokens": 991990, + "tool_calls": 30, + "turns": 31 + }, + "render-026": { + "answer": "dec-ship-four-tern-quality-presets--vjw18d82dx5gath8", + "tokens": 39371, + "fixed_tokens": 1102, + "processed_tokens": 584625, + "tool_calls": 20, + "turns": 21 + }, + "render-027": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 11085, + "fixed_tokens": 1100, + "processed_tokens": 69039, + "tool_calls": 6, + "turns": 7 + }, + "render-028": { + "answer": "dec-move-to-gpu-driven-two-phase-occlusion-culling--6dre481mabxsw6ee", + "tokens": 29492, + "fixed_tokens": 1102, + "processed_tokens": 221212, + "tool_calls": 10, + "turns": 11 + }, + "render-029": { + "answer": "fnd-vsm-page-pool-overflows-during-spring-tides--m2p8e89hhdmh0z8a", + "tokens": 15562, + "fixed_tokens": 1102, + "processed_tokens": 162539, + "tool_calls": 11, + "turns": 12 + }, + "render-030": { + "answer": "dec-use-screen-space-refraction-for-saltglass--vknzksgdrgr7zkg6", + "tokens": 22400, + "fixed_tokens": 1099, + "processed_tokens": 335682, + "tool_calls": 22, + "turns": 21 + }, + "render-031": { + "answer": "Alpha-tested foliage cost 2.1 ms in the visibility buffer (up from 0.8 ms in forward+), because the alpha test runs in the raster pass with a per-pixel texture fetch; saltglass cannot be in the buffer because it refracts the image behind it and needs that resolved image first. See dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9, which derives from fnd-alpha-tested-foliage-costs-2-1-ms-in-the-visibil--f3g34x2k6fs367eb.", + "tokens": 14105, + "fixed_tokens": 1100, + "processed_tokens": 175945, + "tool_calls": 13, + "turns": 14 + }, + "render-032": { + "answer": "Tern was made Lantern's only upscaler because Halcyon has no Skiff build, and a second upscaler would break the one-upscaler constraint; the team also wanted upscaler code it owns for saltglass refraction and caustics, with no vendor license to negotiate. See dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e.", + "tokens": 28128, + "fixed_tokens": 1102, + "processed_tokens": 310138, + "tool_calls": 15, + "turns": 15 + }, + "render-033": { + "answer": "128 MB", + "tokens": 21348, + "fixed_tokens": 1100, + "processed_tokens": 201191, + "tool_calls": 12, + "turns": 13 + }, + "render-034": { + "answer": "PC workstation", + "tokens": 40660, + "fixed_tokens": 1106, + "processed_tokens": 1052303, + "tool_calls": 33, + "turns": 34 + }, + "render-035": { + "answer": "5912 pages", + "tokens": 15735, + "fixed_tokens": 1104, + "processed_tokens": 184355, + "tool_calls": 12, + "turns": 13 + }, + "render-036": { + "answer": "unknown", + "tokens": 100796, + "fixed_tokens": 1101, + "processed_tokens": 5916217, + "tool_calls": 89, + "turns": 90 + }, + "render-037": { + "answer": "unknown", + "tokens": 36381, + "fixed_tokens": 1100, + "processed_tokens": 598844, + "tool_calls": 21, + "turns": 22 + }, + "render-038": { + "answer": "dec-visibility-buffer-for-opaque-geometry-foliage-an--gvnvp04jgm7kbfh9", + "tokens": 26367, + "fixed_tokens": 1104, + "processed_tokens": 259401, + "tool_calls": 13, + "turns": 13 + }, + "render-039": { + "answer": "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "tokens": 42356, + "fixed_tokens": 1103, + "processed_tokens": 463048, + "tool_calls": 16, + "turns": 16 + }, + "render-040": { + "answer": "fnd-lantern-at-alpha--ckpd9ecvdjqbcr7c", + "tokens": 17886, + "fixed_tokens": 1104, + "processed_tokens": 202838, + "tool_calls": 15, + "turns": 16 + }, + "render-041": { + "answer": "4.7 ms", + "tokens": 24485, + "fixed_tokens": 1105, + "processed_tokens": 332505, + "tool_calls": 16, + "turns": 17 + }, + "render-042": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 27383, + "fixed_tokens": 1108, + "processed_tokens": 240807, + "tool_calls": 11, + "turns": 12 + }, + "render-043": { + "answer": "fnd-meshlet-cone-culling-rejects-9-percent-of-triang--t6e5x7trzedex105", + "tokens": 46753, + "fixed_tokens": 1108, + "processed_tokens": 1311511, + "tool_calls": 41, + "turns": 42 + }, + "render-044": { + "answer": "dec-screen-space-refraction-with-a-ray-marched-fallb--2639xtpyrfrvbx1r", + "tokens": 12746, + "fixed_tokens": 1110, + "processed_tokens": 139724, + "tool_calls": 11, + "turns": 12 + }, + "render-045": { + "answer": "false", + "tokens": 26025, + "fixed_tokens": 1099, + "processed_tokens": 203540, + "tool_calls": 10, + "turns": 11 + }, + "render-046": { + "answer": "Virtual shadow maps were reopened because the cost Finding that had rejected them (3.1 ms) was a capture error; with cached pages they cost 1.4 ms, less than four cascades, and cascade seams through saltglass stayed unsolved under CSM. See dec-use-virtual-shadow-maps-for-the-sun--x8nqkn9q7eaf1ap4.", + "tokens": 9924, + "fixed_tokens": 1100, + "processed_tokens": 122519, + "tool_calls": 11, + "turns": 12 + }, + "render-047": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1. Skiff returned to the visibility buffer because the release-build capture measured 7.4 ms against 8.1 ms for forward+ (6.3 ms with half-rate material resolve), so the forward+ fallback was no longer worth keeping.", + "tokens": 23518, + "fixed_tokens": 1104, + "processed_tokens": 307451, + "tool_calls": 17, + "turns": 18 + }, + "render-048": { + "answer": "unknown", + "tokens": 53241, + "fixed_tokens": 1097, + "processed_tokens": 833088, + "tool_calls": 28, + "turns": 25 + }, + "render-049": { + "answer": "unknown", + "tokens": 49466, + "fixed_tokens": 1102, + "processed_tokens": 1119452, + "tool_calls": 33, + "turns": 34 + }, + "render-050": { + "answer": "con-frame-budget--fzdrbxns6dmrrmy3", + "tokens": 9903, + "fixed_tokens": 1098, + "processed_tokens": 57217, + "tool_calls": 5, + "turns": 6 + }, + "render-051": { + "answer": "iss-tern-ghosting-on-moving-saltglass--2yr3q6zf476wajqp", + "tokens": 31258, + "fixed_tokens": 1104, + "processed_tokens": 293271, + "tool_calls": 12, + "turns": 13 + }, + "render-052": { + "answer": "61 percent", + "tokens": 38023, + "fixed_tokens": 1103, + "processed_tokens": 708480, + "tool_calls": 25, + "turns": 26 + }, + "render-053": { + "answer": "every frame", + "tokens": 40669, + "fixed_tokens": 1103, + "processed_tokens": 507180, + "tool_calls": 17, + "turns": 18 + }, + "render-054": { + "answer": "debug-gpu-validation", + "tokens": 27965, + "fixed_tokens": 1102, + "processed_tokens": 523144, + "tool_calls": 21, + "turns": 22 + }, + "render-055": { + "answer": "fnd-tern-cost-at-4k-output-on-meridian-with-the-reac--81tcpzvhwzcy6wyn", + "tokens": 39953, + "fixed_tokens": 1106, + "processed_tokens": 535484, + "tool_calls": 19, + "turns": 20 + }, + "render-056": { + "answer": "208,000 meshlets", + "tokens": 28177, + "fixed_tokens": 1102, + "processed_tokens": 414771, + "tool_calls": 17, + "turns": 18 + }, + "render-057": { + "answer": "dec-visibility-buffer-on-all-platforms-skiff-resolve--k1ackk3c3275r5y1", + "tokens": 29679, + "fixed_tokens": 1100, + "processed_tokens": 241184, + "tool_calls": 12, + "turns": 12 + }, + "render-058": { + "answer": "iss-which-upscaler-should-lantern-ship--g8c2y9tsr856tn2v", + "tokens": 42668, + "fixed_tokens": 1105, + "processed_tokens": 1061000, + "tool_calls": 33, + "turns": 34 + }, + "render-059": { + "answer": "fnd-sdf-gi-costs-4-7-ms-on-meridian--sm19rt2f274ajsvr", + "tokens": 40442, + "fixed_tokens": 1107, + "processed_tokens": 708290, + "tool_calls": 23, + "turns": 24 + }, + "render-060": { + "answer": "0.31", + "tokens": 29575, + "fixed_tokens": 1100, + "processed_tokens": 399821, + "tool_calls": 16, + "turns": 17 + }, + "render-061": { + "answer": "con-ui-composites-at-a-fixed-203-nits-in-hdr--de6aft7cvhdezn5s", + "tokens": 33143, + "fixed_tokens": 1105, + "processed_tokens": 429548, + "tool_calls": 16, + "turns": 17 + }, + "render-062": { + "answer": "con-frame-budget--j6vneva1g2ywj86m", + "tokens": 29883, + "fixed_tokens": 1102, + "processed_tokens": 389284, + "tool_calls": 16, + "turns": 17 + }, + "render-063": { + "answer": "dec-resolve-materials-with-one-uber-shader-over-bind--s9nbq2qn05mk5j6r", + "tokens": 25408, + "fixed_tokens": 1101, + "processed_tokens": 182611, + "tool_calls": 11, + "turns": 10 + }, + "render-064": { + "answer": "dec-use-a-cached-shadow-atlas-for-local-lights--8p6f2v774he38t5j", + "tokens": 42779, + "fixed_tokens": 1103, + "processed_tokens": 403307, + "tool_calls": 14, + "turns": 14 + }, + "render-065": { + "answer": "dec-ship-tern-as-lantern-s-only-upscaler--fmreq35nzyjc7k4e", + "tokens": 61233, + "fixed_tokens": 1100, + "processed_tokens": 2004215, + "tool_calls": 49, + "turns": 50 + }, + "render-066": { + "answer": "dec-use-cascaded-shadow-maps-with-four-cascades--s5c1y9jaf275afzr", + "tokens": 31397, + "fixed_tokens": 1105, + "processed_tokens": 431929, + "tool_calls": 18, + "turns": 19 + }, + "render-067": { + "answer": 4, + "tokens": 15579, + "fixed_tokens": 1104, + "processed_tokens": 213242, + "tool_calls": 15, + "turns": 16 + }, + "render-068": { + "answer": [ + "dec-exclude-caustic-luminance-from-tern-s-history-cl--xhxjn26ss343v2k4", + "dec-lock-caustic-photon-jitter-to-the-tide-phase--85x10k4h11t8nx7c", + "dec-use-the-probe-grid-alone-in-120-hz-performance-m--6dzm1nt40pjfpyae", + "dec-trace-sdf-gi-at-quarter-resolution-in-performanc--06cqt3mqn3fg5h6e", + "dec-read-paper-white-from-the-meridian-system-settin--7crbx73gqw1hn7xq", + "dec-add-an-in-game-paper-white-slider-from-80-to-400--1phfvxby983z8tjf" + ], + "tokens": 23661, + "fixed_tokens": 1107, + "processed_tokens": 152362, + "tool_calls": 8, + "turns": 9 + }, + "render-069": { + "answer": "iss-shadow-pages-run-out-at-spring-low-tide--f4fgy6b44jbkrsmh", + "tokens": 18488, + "fixed_tokens": 1098, + "processed_tokens": 177433, + "tool_calls": 12, + "turns": 13 + }, + "render-070": { + "answer": "unknown", + "tokens": 63542, + "fixed_tokens": 1103, + "processed_tokens": 1605955, + "tool_calls": 39, + "turns": 38 + }, + "render-071": { + "answer": 3, + "tokens": 10680, + "fixed_tokens": 1100, + "processed_tokens": 72219, + "tool_calls": 8, + "turns": 7 + }, + "render-072": { + "answer": [ + "rsk-sdf-volumes-outgrow-the-render-target-memory-bud--xfd2pfztvskdcd2c", + "rsk-artists-cannot-match-concept-paintings-under-the--6cfh9vkamn0kbtcx" + ], + "tokens": 12054, + "fixed_tokens": 1105, + "processed_tokens": 89249, + "tool_calls": 9, + "turns": 8 + }, + "render-073": { + "answer": [ + "rsk-vsm-page-pool-overflows-when-tides-expose-the-fl--x1r161nrdngra4ty", + "rsk-gi-cannot-fit-the-120-hz-performance-mode--tqt3fj839dqfq5s7", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 13577, + "fixed_tokens": 1108, + "processed_tokens": 107641, + "tool_calls": 11, + "turns": 9 + }, + "render-074": { + "answer": 5, + "tokens": 7442, + "fixed_tokens": 1100, + "processed_tokens": 52812, + "tool_calls": 5, + "turns": 6 + }, + "render-075": { + "answer": "3", + "tokens": 36047, + "fixed_tokens": 1099, + "processed_tokens": 1005869, + "tool_calls": 34, + "turns": 35 + }, + "render-076": { + "answer": false, + "tokens": 18889, + "fixed_tokens": 1103, + "processed_tokens": 233898, + "tool_calls": 18, + "turns": 16 + }, + "render-077": { + "answer": true, + "tokens": 23270, + "fixed_tokens": 1110, + "processed_tokens": 369837, + "tool_calls": 20, + "turns": 21 + }, + "render-078": { + "answer": [ + "dec-brine-curve-with-a-separate-hdr-shoulder-at-1-00--be68yywg78dysm2e", + "rsk-hdr-certification-slips-the-release-candidate--49t4xjbzmkgc22a4" + ], + "tokens": 36454, + "fixed_tokens": 1107, + "processed_tokens": 373192, + "tool_calls": 16, + "turns": 15 + }, + "render-079": { + "answer": 5, + "tokens": 55681, + "fixed_tokens": 1099, + "processed_tokens": 501637, + "tool_calls": 18, + "turns": 17 + }, + "render-080": { + "answer": "6 versions", + "tokens": 13195, + "fixed_tokens": 1098, + "processed_tokens": 117366, + "tool_calls": 9, + "turns": 10 + }, + "render-081": { + "answer": "unknown", + "tokens": 41195, + "fixed_tokens": 1103, + "processed_tokens": 1028297, + "tool_calls": 33, + "turns": 34 + }, + "render-082": { + "answer": "false", + "tokens": 25858, + "fixed_tokens": 1101, + "processed_tokens": 515864, + "tool_calls": 27, + "turns": 28 + }, + "render-083": { + "answer": null, + "tokens": 71970, + "fixed_tokens": 1100, + "processed_tokens": 1815091, + "tool_calls": 47, + "turns": 47 + }, + "tools-001": { + "answer": "completed", + "tokens": 18740, + "fixed_tokens": 1104, + "processed_tokens": 130158, + "tool_calls": 8, + "turns": 9 + }, + "tools-002": { + "answer": "47 minutes", + "tokens": 19509, + "fixed_tokens": 1104, + "processed_tokens": 164029, + "tool_calls": 9, + "turns": 10 + }, + "tools-003": { + "answer": "6 min 40 s", + "tokens": 37828, + "fixed_tokens": 1106, + "processed_tokens": 855388, + "tool_calls": 29, + "turns": 30 + }, + "tools-004": { + "answer": "unknown", + "tokens": 42304, + "fixed_tokens": 1102, + "processed_tokens": 371858, + "tool_calls": 12, + "turns": 13 + }, + "tools-005": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 22822, + "fixed_tokens": 1103, + "processed_tokens": 197312, + "tool_calls": 10, + "turns": 11 + }, + "tools-006": { + "answer": "dec-build-a-content-addressed-shader-compile-cache-s--e65rq9439k7xz5zh", + "tokens": 8530, + "fixed_tokens": 1103, + "processed_tokens": 86773, + "tool_calls": 8, + "turns": 9 + }, + "tools-007": { + "answer": "dec-raise-the-shader-cache-fetch-timeout-from-5-seco--4sgdf42ajyct6gt6", + "tokens": 56087, + "fixed_tokens": 1109, + "processed_tokens": 1156192, + "tool_calls": 31, + "turns": 29 + }, + "tools-008": { + "answer": "dec-key-the-shader-cache-on-raw-source-text-and-comp--hp6awmx3gq5rn9nc", + "tokens": 37816, + "fixed_tokens": 1104, + "processed_tokens": 349746, + "tool_calls": 15, + "turns": 13 + }, + "tools-009": { + "answer": "92.4%", + "tokens": 27391, + "fixed_tokens": 1102, + "processed_tokens": 477242, + "tool_calls": 20, + "turns": 21 + }, + "tools-010": { + "answer": "dec-cap-each-farm-agent-at-6-parallel-shader-compile--q5w4sfjm73tsj8x8", + "tokens": 29152, + "fixed_tokens": 1104, + "processed_tokens": 477211, + "tool_calls": 20, + "turns": 21 + }, + "tools-011": { + "answer": "fnd-two-compiler-builds-with-the-same-version-string--gmgr4ha0ndxqmcn9", + "tokens": 36811, + "fixed_tokens": 1102, + "processed_tokens": 324810, + "tool_calls": 12, + "turns": 13 + }, + "tools-012": { + "answer": 3, + "tokens": 22005, + "fixed_tokens": 1107, + "processed_tokens": 167671, + "tool_calls": 9, + "turns": 10 + }, + "tools-013": { + "answer": 8, + "tokens": 86412, + "fixed_tokens": 1105, + "processed_tokens": 2278248, + "tool_calls": 47, + "turns": 48 + }, + "tools-014": { + "answer": false, + "tokens": 22989, + "fixed_tokens": 1105, + "processed_tokens": 245682, + "tool_calls": 12, + "turns": 13 + }, + "tools-015": { + "answer": [ + "rsk-shader-cache-storage-could-outgrow-the-4-tb-farm--76y9649mwrx7aavd", + "iss-whether-the-audio-bank-packer-should-load-on-a-w--sqzp2rwp6s01w3b0", + "iss-whether-the-level-editor-autosave-needs-its-own--rkvmnbqe7w644zkr", + "iss-whether-the-artist-sandbox-should-load-on-a-work--73j819anqct8etfn", + "iss-whether-the-review-bot-should-load-on-a-worker-t--rtsjwvsksjdq15v1", + "iss-whether-the-level-editor-autosave-should-be-cach--radcjs4hyvh7p55q", + "iss-whether-the-crash-symbol-server-should-load-on-a--qvyphhgnsnm2cvqb", + "iss-whether-the-license-scanner-should-be-cached-bet--nqhd89czbfat9rya", + "iss-whether-the-doc-site-should-load-on-a-worker-thr--qxswe0pg52ekky0z", + "iss-whether-the-asset-search-should-load-on-a-worker--f9gpv3z522x6p6h4", + "iss-whether-the-changelog-bot-should-load-on-a-worke--jgtdn2f02p8fx5nr", + "iss-whether-the-review-bot-needs-its-own-memory-pool--pg3m8g9gn3y6v3kq", + "iss-whether-the-changelog-bot-should-skip-work-when--p2vrp5d682p3jrc3", + "iss-whether-the-patch-differ-should-load-on-a-worker--cysv4n56q1qn7t2k", + "iss-whether-the-artist-sandbox-should-skip-work-when--11tqzws72gms2h8x", + "iss-whether-the-audio-bank-packer-needs-its-own-memo--qjrqrseast8d5mhf", + "iss-whether-the-changelog-bot-needs-its-own-memory-p--mw8zejz6hk5m60x0", + "iss-whether-the-crash-symbol-server-should-ship-in-t--bhq91ajavah18wsk", + "iss-whether-the-asset-search-should-load-on-a-worker--80b27kqh36rpafdt", + "iss-whether-the-doc-site-should-ship-in-the-first-pa--wcdkq5zab7dxe5yj", + "iss-whether-the-patch-differ-should-be-cached-betwee--28b5zqb1ketwzf0r", + "iss-whether-the-test-runner-should-load-on-a-worker--aygay1r0qj4mzf78", + "iss-whether-the-texture-viewer-should-be-split-into--1jhgyqb70h5x7m7b", + "iss-whether-the-level-editor-autosave-should-skip-wo--fyysrcv6fvbsvcsy", + "iss-whether-the-level-editor-autosave-should-be-spli--ncnwh0csw671fs0q", + "iss-whether-the-asset-search-should-ship-in-the-firs--hsb2avhx613prxew", + "iss-whether-the-test-runner-needs-its-own-memory-poo--bs6arq0qv37d6wmx", + "iss-whether-the-artist-sandbox-should-be-split-into--ek47ca2vgmrjanp8" + ], + "tokens": 77670, + "fixed_tokens": 1105, + "processed_tokens": 800342, + "tool_calls": 20, + "turns": 20 + }, + "tools-016": { + "answer": "dec-key-the-shader-cache-on-preprocessed-source-defi--qx2da8e4cbkx0drp", + "tokens": 41714, + "fixed_tokens": 1100, + "processed_tokens": 924803, + "tool_calls": 30, + "turns": 31 + }, + "world-001": { + "answer": "64 KB", + "tokens": 15736, + "fixed_tokens": 1103, + "processed_tokens": 230538, + "tool_calls": 16, + "turns": 17 + }, + "world-002": { + "answer": "5.1 ms", + "tokens": 35377, + "fixed_tokens": 1102, + "processed_tokens": 400782, + "tool_calls": 14, + "turns": 15 + }, + "world-003": { + "answer": "1 crash per 900 script waits", + "tokens": 36620, + "fixed_tokens": 1099, + "processed_tokens": 508893, + "tool_calls": 18, + "turns": 19 + }, + "world-004": { + "answer": "4,100 file opens", + "tokens": 38565, + "fixed_tokens": 1103, + "processed_tokens": 556345, + "tool_calls": 20, + "turns": 21 + }, + "world-005": { + "answer": "11 m", + "tokens": 110334, + "fixed_tokens": 1099, + "processed_tokens": 1833939, + "tool_calls": 28, + "turns": 27 + }, + "world-006": { + "answer": "3.4 reversals per minute", + "tokens": 32232, + "fixed_tokens": 1099, + "processed_tokens": 462461, + "tool_calls": 19, + "turns": 19 + }, + "world-007": { + "answer": "unknown", + "tokens": 52614, + "fixed_tokens": 1102, + "processed_tokens": 1303320, + "tool_calls": 35, + "turns": 36 + }, + "world-008": { + "answer": "10 Hz", + "tokens": 36741, + "fixed_tokens": 1101, + "processed_tokens": 576021, + "tool_calls": 20, + "turns": 21 + }, + "world-009": { + "answer": "unknown", + "tokens": 34701, + "fixed_tokens": 1102, + "processed_tokens": 698208, + "tool_calls": 25, + "turns": 26 + }, + "world-010": { + "answer": "29 GB/s", + "tokens": 13434, + "fixed_tokens": 1106, + "processed_tokens": 107424, + "tool_calls": 8, + "turns": 9 + }, + "world-011": { + "answer": "fnd-shipping-entity-names-costs-57-mb-at-full-popula--6w7gjvy7wdmdx9qx", + "tokens": 64703, + "fixed_tokens": 1103, + "processed_tokens": 2747688, + "tool_calls": 57, + "turns": 58 + }, + "world-012": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 22016, + "fixed_tokens": 1102, + "processed_tokens": 140827, + "tool_calls": 8, + "turns": 9 + }, + "world-013": { + "answer": "unknown", + "tokens": 66373, + "fixed_tokens": 1102, + "processed_tokens": 2018146, + "tool_calls": 43, + "turns": 44 + }, + "world-014": { + "answer": "unknown", + "tokens": 39184, + "fixed_tokens": 1103, + "processed_tokens": 815757, + "tool_calls": 26, + "turns": 27 + }, + "world-015": { + "answer": "3.0 ms", + "tokens": 78563, + "fixed_tokens": 1099, + "processed_tokens": 968322, + "tool_calls": 19, + "turns": 20 + }, + "world-016": { + "answer": "unknown", + "tokens": 21085, + "fixed_tokens": 1097, + "processed_tokens": 138728, + "tool_calls": 8, + "turns": 9 + }, + "world-017": { + "answer": "dec-partition-the-world-into-128-m-cells--372b6czbjf29ymaw", + "tokens": 31191, + "fixed_tokens": 1102, + "processed_tokens": 375931, + "tool_calls": 17, + "turns": 16 + }, + "world-018": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 22736, + "fixed_tokens": 1106, + "processed_tokens": 156044, + "tool_calls": 9, + "turns": 10 + }, + "world-019": { + "answer": "fnd-tide-priority-reads-starve-player-proximity-load--by0d7b0bk39xz3rk", + "tokens": 48296, + "fixed_tokens": 1107, + "processed_tokens": 732659, + "tool_calls": 21, + "turns": 22 + }, + "world-020": { + "answer": "dec-schedule-streaming-reads-by-deadline-instead-of--q2x5aknabgcssdj2", + "tokens": 12002, + "fixed_tokens": 1102, + "processed_tokens": 99283, + "tool_calls": 8, + "turns": 9 + }, + "world-021": { + "answer": "dec-cut-navigation-tiles-to-64-m-so-three-fit-along--snvepk8ygeb3qsjs", + "tokens": 28593, + "fixed_tokens": 1103, + "processed_tokens": 749643, + "tool_calls": 37, + "turns": 38 + }, + "world-022": { + "answer": "fnd-41-of-alpha-crash-reports-have-no-usable-stack--35a50b07ferhf13m", + "tokens": 16620, + "fixed_tokens": 1099, + "processed_tokens": 164623, + "tool_calls": 11, + "turns": 12 + }, + "world-023": { + "answer": "dec-compress-streamed-assets-with-kelp--ezwc5p1tgsgm7v8m", + "tokens": 34782, + "fixed_tokens": 1100, + "processed_tokens": 771488, + "tool_calls": 31, + "turns": 32 + }, + "world-024": { + "answer": "dec-shrink-kelp-s-dictionary-to-1-mb-for-streamed-as--yh8hktfqnyqhepew", + "tokens": 32491, + "fixed_tokens": 1105, + "processed_tokens": 466568, + "tool_calls": 18, + "turns": 19 + }, + "world-025": { + "answer": "con-streaming-pool-budget--kca88sgh2j6sxvd5", + "tokens": 23567, + "fixed_tokens": 1099, + "processed_tokens": 136157, + "tool_calls": 8, + "turns": 9 + }, + "world-026": { + "answer": "fnd-retail-skiff-units-sustain-only-190-mb-s-of-stre--485sdcmm3gweevvn", + "tokens": 54697, + "fixed_tokens": 1105, + "processed_tokens": 1442935, + "tool_calls": 38, + "turns": 39 + }, + "world-027": { + "answer": "160 MB/s", + "tokens": 38420, + "fixed_tokens": 1099, + "processed_tokens": 280465, + "tool_calls": 12, + "turns": 11 + }, + "world-028": { + "answer": "unknown", + "tokens": 68168, + "fixed_tokens": 1098, + "processed_tokens": 1666708, + "tool_calls": 35, + "turns": 34 + }, + "world-029": { + "answer": "dec-run-engine-jobs-and-gull-scripts-on-one-thread-p--07pbzvqg00pmnrr3", + "tokens": 19934, + "fixed_tokens": 1100, + "processed_tokens": 270564, + "tool_calls": 17, + "turns": 18 + }, + "world-030": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 26660, + "fixed_tokens": 1104, + "processed_tokens": 209319, + "tool_calls": 10, + "turns": 11 + }, + "world-031": { + "answer": "unknown", + "tokens": 88449, + "fixed_tokens": 1106, + "processed_tokens": 1591120, + "tool_calls": 26, + "turns": 27 + }, + "world-032": { + "answer": "fnd-float-tide-heights-drift-apart-between-pc-and-me--7jb0fg4b5xzr7q5e", + "tokens": 30210, + "fixed_tokens": 1101, + "processed_tokens": 694230, + "tool_calls": 34, + "turns": 35 + }, + "world-033": { + "answer": "dec-save-player-changes-as-deltas-against-the-tide-f--n38scy1yaarsr8cw", + "tokens": 39474, + "fixed_tokens": 1102, + "processed_tokens": 669491, + "tool_calls": 22, + "turns": 23 + }, + "world-034": { + "answer": "unknown", + "tokens": 38692, + "fixed_tokens": 1099, + "processed_tokens": 1383990, + "tool_calls": 57, + "turns": 58 + }, + "world-035": { + "answer": false, + "tokens": 28444, + "fixed_tokens": 1102, + "processed_tokens": 324950, + "tool_calls": 14, + "turns": 15 + }, + "world-036": { + "answer": "Kestrel retired the fiber scheduler because 41% of alpha crash reports had no usable stack (fiber workers showed only scheduler frames), which blocked crash triage for beta, while a continuation-based thread pool kept workers 93% busy versus 95% with fibers. See dec-return-engine-jobs-to-a-thread-pool-with-continu--1zkc1v2x97taf50h.", + "tokens": 26550, + "fixed_tokens": 1100, + "processed_tokens": 319896, + "tool_calls": 14, + "turns": 15 + }, + "world-037": { + "answer": "fnd-hot-reload-drops-suspended-gull-coroutines--va0fn0svryzwbzws", + "tokens": 49125, + "fixed_tokens": 1103, + "processed_tokens": 930867, + "tool_calls": 27, + "turns": 28 + }, + "world-038": { + "answer": "3", + "tokens": 26948, + "fixed_tokens": 1103, + "processed_tokens": 369242, + "tool_calls": 16, + "turns": 17 + }, + "world-039": { + "answer": "unknown", + "tokens": 44141, + "fixed_tokens": 1103, + "processed_tokens": 1030695, + "tool_calls": 30, + "turns": 31 + }, + "world-040": { + "answer": "unknown", + "tokens": 55594, + "fixed_tokens": 1103, + "processed_tokens": 946996, + "tool_calls": 22, + "turns": 23 + }, + "world-041": { + "answer": "61%", + "tokens": 25441, + "fixed_tokens": 1102, + "processed_tokens": 231834, + "tool_calls": 11, + "turns": 12 + }, + "world-042": { + "answer": "5.4 s", + "tokens": 30231, + "fixed_tokens": 1107, + "processed_tokens": 349424, + "tool_calls": 14, + "turns": 15 + }, + "world-043": { + "answer": "5.8 minutes", + "tokens": 37695, + "fixed_tokens": 1102, + "processed_tokens": 604405, + "tool_calls": 21, + "turns": 22 + }, + "world-044": { + "answer": "dec-move-engine-jobs-to-fibers--qjkgnm8cex8ydzfq", + "tokens": 28664, + "fixed_tokens": 1104, + "processed_tokens": 348034, + "tool_calls": 15, + "turns": 16 + }, + "world-045": { + "answer": "unknown", + "tokens": 32167, + "fixed_tokens": 1101, + "processed_tokens": 462901, + "tool_calls": 18, + "turns": 19 + }, + "world-046": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 24833, + "fixed_tokens": 1104, + "processed_tokens": 183471, + "tool_calls": 9, + "turns": 10 + }, + "world-047": { + "answer": "fnd-world-core-state-at-alpha--tkbsm4867h65jqq8", + "tokens": 16321, + "fixed_tokens": 1103, + "processed_tokens": 145818, + "tool_calls": 11, + "turns": 12 + }, + "world-048": { + "answer": "dec-compile-gull-to-bytecode-for-a-register-vm--jrqje646fpevp5f3", + "tokens": 33072, + "fixed_tokens": 1106, + "processed_tokens": 409769, + "tool_calls": 17, + "turns": 18 + }, + "world-049": { + "answer": "22%", + "tokens": 44868, + "fixed_tokens": 1099, + "processed_tokens": 1088188, + "tool_calls": 32, + "turns": 33 + }, + "world-050": { + "answer": "fnd-alpha-content-needs-6-2-ms-of-gull-time-per-fram--4mekgdpsdv5nk085", + "tokens": 35269, + "fixed_tokens": 1104, + "processed_tokens": 488213, + "tool_calls": 18, + "turns": 19 + }, + "world-051": { + "answer": "unknown", + "tokens": 37982, + "fixed_tokens": 1105, + "processed_tokens": 649071, + "tool_calls": 22, + "turns": 23 + }, + "world-052": { + "answer": "rsk-ports-may-need-to-shrink-the-streaming-pool-for--1gtjkygrtkb5kgww", + "tokens": 17718, + "fixed_tokens": 1102, + "processed_tokens": 82912, + "tool_calls": 8, + "turns": 7 + }, + "world-053": { + "answer": "unknown", + "tokens": 56666, + "fixed_tokens": 1106, + "processed_tokens": 1659826, + "tool_calls": 38, + "turns": 39 + }, + "world-054": { + "answer": "Marta Lindqvist", + "tokens": 28527, + "fixed_tokens": 1103, + "processed_tokens": 381818, + "tool_calls": 16, + "turns": 17 + }, + "world-055": { + "answer": "unknown", + "tokens": 55694, + "fixed_tokens": 1105, + "processed_tokens": 783873, + "tool_calls": 23, + "turns": 22 + }, + "world-056": { + "answer": "dec-run-engine-jobs-on-a-work-stealing-thread-pool--1pd72x1pxpgwy21x", + "tokens": 23679, + "fixed_tokens": 1110, + "processed_tokens": 197430, + "tool_calls": 10, + "turns": 11 + }, + "world-057": { + "answer": "tile [212, 87]", + "tokens": 17806, + "fixed_tokens": 1107, + "processed_tokens": 259791, + "tool_calls": 16, + "turns": 17 + }, + "world-058": { + "answer": "0.8 GB/s", + "tokens": 39832, + "fixed_tokens": 1104, + "processed_tokens": 872304, + "tool_calls": 29, + "turns": 30 + }, + "world-059": { + "answer": "dec-derive-the-missing-harmonic-phase-from-each-save--gbqrvvgyq2gy2gnw", + "tokens": 56547, + "fixed_tokens": 1106, + "processed_tokens": 1017748, + "tool_calls": 34, + "turns": 29 + }, + "world-060": { + "answer": "dec-partition-the-world-into-192-m-cells--bk9r1c5jwb6fh8tc", + "tokens": 15560, + "fixed_tokens": 1106, + "processed_tokens": 183621, + "tool_calls": 12, + "turns": 13 + }, + "world-061": { + "answer": "236 MB", + "tokens": 31231, + "fixed_tokens": 1106, + "processed_tokens": 414561, + "tool_calls": 16, + "turns": 17 + }, + "world-062": { + "answer": "310 MB", + "tokens": 15984, + "fixed_tokens": 1098, + "processed_tokens": 274154, + "tool_calls": 19, + "turns": 20 + }, + "world-063": { + "answer": "con-streaming-pool-budget--632bkbs79bwj4cr5", + "tokens": 33710, + "fixed_tokens": 1107, + "processed_tokens": 540934, + "tool_calls": 21, + "turns": 22 + }, + "world-064": { + "answer": [ + "dec-force-strict-ieee-float-mode-for-the-tide-simula--ecwmqt7ncjngysqw", + "dec-have-the-host-stream-tide-heights-to-every-peer--a3rb1seedn2bj7k8" + ], + "tokens": 29155, + "fixed_tokens": 1100, + "processed_tokens": 337749, + "tool_calls": 17, + "turns": 18 + }, + "world-065": { + "answer": "3.1 ms", + "tokens": 18303, + "fixed_tokens": 1106, + "processed_tokens": 366567, + "tool_calls": 24, + "turns": 25 + }, + "world-066": { + "answer": "46", + "tokens": 21181, + "fixed_tokens": 1104, + "processed_tokens": 338050, + "tool_calls": 19, + "turns": 20 + }, + "world-067": { + "answer": "1.2 GB/s", + "tokens": 23756, + "fixed_tokens": 1098, + "processed_tokens": 328636, + "tool_calls": 17, + "turns": 18 + }, + "world-068": { + "answer": true, + "tokens": 37116, + "fixed_tokens": 1100, + "processed_tokens": 743212, + "tool_calls": 31, + "turns": 32 + }, + "world-069": { + "answer": 6, + "tokens": 52452, + "fixed_tokens": 1103, + "processed_tokens": 954802, + "tool_calls": 28, + "turns": 29 + }, + "world-070": { + "answer": 3, + "tokens": 19295, + "fixed_tokens": 1101, + "processed_tokens": 127769, + "tool_calls": 8, + "turns": 9 + }, + "world-071": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--yxsjd26wcaz6fgx9" + ], + "tokens": 40159, + "fixed_tokens": 1103, + "processed_tokens": 404081, + "tool_calls": 16, + "turns": 15 + }, + "world-072": { + "answer": [ + "dec-compress-streamed-assets-with-zstandard--bs1t15cygc51cpje", + "dec-compress-streamed-assets-with-lz4--v1pfsgfp0b3yjv4s" + ], + "tokens": 48794, + "fixed_tokens": 1100, + "processed_tokens": 505729, + "tool_calls": 19, + "turns": 18 + }, + "world-073": { + "answer": false, + "tokens": 42788, + "fixed_tokens": 1110, + "processed_tokens": 414136, + "tool_calls": 16, + "turns": 16 + }, + "world-074": { + "answer": 7, + "tokens": 9171, + "fixed_tokens": 1103, + "processed_tokens": 69313, + "tool_calls": 6, + "turns": 7 + }, + "world-075": { + "answer": [ + "dec-store-components-in-archetype-tables-and-volatil--8679sjn071ycsbpv", + "dec-partition-the-world-into-192-meter-cells--q62ddwdv2n8wxay6", + "dec-return-engine-jobs-to-the-thread-pool-with-conti--547qbweqvrp1j5g3" + ], + "tokens": 6273, + "fixed_tokens": 1103, + "processed_tokens": 50135, + "tool_calls": 5, + "turns": 6 + }, + "world-076": { + "answer": "unknown", + "tokens": 94556, + "fixed_tokens": 1107, + "processed_tokens": 2291807, + "tool_calls": 70, + "turns": 40 + }, + "world-077": { + "answer": 251, + "tokens": 27284, + "fixed_tokens": 1102, + "processed_tokens": 369240, + "tool_calls": 20, + "turns": 21 + }, + "world-078": { + "answer": [ + "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "rsk-reworking-the-read-scheduler-may-slip-the-coast--9jzvs7etvb5jxqtp" + ], + "tokens": 12720, + "fixed_tokens": 1106, + "processed_tokens": 85210, + "tool_calls": 7, + "turns": 8 + }, + "world-079": { + "answer": [ + "con-no-runtime-code-generation-on-consoles--6cf25zqmsvqkdd73", + "con-tide-state-must-be-bit-identical-on-every-peer--zrpnp8qjyxzkas7t", + "con-saves-from-beta-onward-must-load-in-every-later--v6z856sxtgc5sb09", + "con-streaming-pool-budget--8czwb4fjr5p5vpy7" + ], + "tokens": 16567, + "fixed_tokens": 1102, + "processed_tokens": 115954, + "tool_calls": 14, + "turns": 9 + }, + "world-080": { + "answer": [ + "iss-gull-hot-reload-rejects-struct-layout-changes--3v2w2f9hg2gwy6aj", + "iss-gull-hot-reload-drops-suspended-coroutines--z7j2gmyny6gdxd07", + "iss-cross-platform-cloud-saves-are-unplanned--5bk0k4wrgj42ztfy" + ], + "tokens": 34441, + "fixed_tokens": 1103, + "processed_tokens": 275281, + "tool_calls": 11, + "turns": 12 + }, + "world-081": { + "answer": "rsk-hot-reload-may-corrupt-live-script-state--ec2ht2r1sq6h56ew", + "tokens": 58826, + "fixed_tokens": 1098, + "processed_tokens": 1702797, + "tool_calls": 42, + "turns": 43 + } + } +}