[onkarbpatil]

Summary

Adds OpenInfer (openinfer.io) as an inference provider, with mappings for the two models it currently serves. Both models are new to the registry, so their canonical model entries are included.

Disclosure: I work at OpenInfer, so provider details (model catalog, rate limits) come from first-hand knowledge of the service.

Files added

• providers/openinfer.yaml — OpenAI-compatible API at https://api.openinfer.ai/v1, bearer auth, global routing across distributed CPU/GPU/NPU compute
• models/qwen3-5-9b.yaml — Qwen 3.5 9B (Alibaba, Apache 2.0, 262K context, multimodal)
• models/llama-3-2-1b.yaml — Llama 3.2 1B Instruct (Meta, 128K context)
• mappings/openinfer/qwen3-5-9b.yaml — served as @oi/Qwen3.5-9B
• mappings/openinfer/llama-3-2-1b.yaml — served as @oi/Llama-3.2-1B-Instruct

Pricing note

OpenInfer Cloud is free through August 31, 2026 (see openinfer.io/cloud), so mappings list $0.00 input/output. We will submit an update with post-promo rates when the free period ends.

Validation

python validate_registry.py passes — 403 files validated.

🤖 Generated with Claude Code

[onkarbpatil]

Hi, I ahve added this PR to integrate OpenInfer's OICloud as an infernece provider service. I have the API key generated, but wasn't sure where to add or how to share. Please help with that.

[onkarbpatil]

Hi, @madeburo, could we get an update on merging this PR?

[madeburo]

Hi @onkarbpatil
Thanks for the contribution the registry structure looks good and local validation passes.

Before we can merge, please address the following:

1. The mappings currently list $0.00 pricing based on the free preview ending August 31, 2026. Since that date has passed, please confirm that the promotion was extended or update both mappings with the current post-preview pricing and an official public source.

2. Our provider requirements specify a publicly accessible API rather than an invite-only/private preview. The OpenInfer Cloud page currently shows “free preview” and “Request Access.” Please clarify the current access model and provide a public signup or documentation link.

3. Please provide public references for the model IDs, 60 RPM / 100K TPM limits, and provider-supported context windows. If OpenInfer supports a smaller context than the canonical models, please set context_window_override accordingly.

Please do not share the API key in this PR or commit it to the repository. API credentials can be handled separately through a private channel or repository secrets when telemetry integration is configured.
Once these points are updated, we’ll be happy to review the PR again. Thank you!

[onkarbpatil]

Hi @madeburo I have added the pricing and context size. You can get the public API to access them here "https://console.openinfer.ai/usage"

[onkarbpatil]

Hi @madeburo pinging for update.

[onkarbpatil]

@madeburo any update?

[onkarbpatil]

@madeburo pinging for update again.

[chongyihu]

@madeburo
Sorry for the repeated pings! Could you help us get this reviewed and approved? Let me know if anything's missing or if there's anything I can do to help move it along.

[madeburo]

Hi @onkarbpatil, @chongyihu — thanks for the updates,

The files validate against the current main and the model entries look correct. Before we can merge the provider, we still need publicly verifiable sources:

1. Pricing: console.openinfer.ai/usage requires a login, so we can't verify the per-token prices. Please link a public pricing page or public docs.
2. Access: openinfer.io/cloud still shows a "Get Access" form and "free through August 31, 2026". The registry only lists providers with public self-serve signup. Please share the public signup link once it's available.
3. Limits and model IDs: please link public documentation for the @oi/... model IDs, the 60 RPM / 100K TPM limits, and the served context windows.

Please also update the PR description, which still mentions $0.00 pricing and two models.

If it helps, we can merge the canonical model entries (Qwen 3.5 9B, Qwen 3.5 27B, Llama 3.2 1B) separately in the meantime. Just let us know.

[chongyihu]

Hi @onkarbpatil, @chongyihu — thanks for the updates,

The files validate against the current main and the model entries look correct. Before we can merge the provider, we still need publicly verifiable sources:

1. Pricing: console.openinfer.ai/usage requires a login, so we can't verify the per-token prices. Please link a public pricing page or public docs.
2. Access: openinfer.io/cloud still shows a "Get Access" form and "free through August 31, 2026". The registry only lists providers with public self-serve signup. Please share the public signup link once it's available.
3. Limits and model IDs: please link public documentation for the @oi/... model IDs, the 60 RPM / 100K TPM limits, and the served context windows.

Please also update the PR description, which still mentions $0.00 pricing and two models.

If it helps, we can merge the canonical model entries (Qwen 3.5 9B, Qwen 3.5 27B, Llama 3.2 1B) separately in the meantime. Just let us know.

@madeburo Thank you! I will report back and let my team revise this. I will get back to you ASAP.

[cbuffington-cell]

Hi @madeburo
Hope all is well. Can you please let me and @chongyihu know if there is anything else you need? We are ready on our end to complete this merge and get up and running. Thank you!

[madeburo]

Thanks for following up. I re-reviewed the current head and tested a merge with the latest main: there were no conflicts, and local registry validation passed for all 587 files. The pricing and served-context overrides added in September are present. The remaining issues are public verification and consistency of the submission:

1. Public access. The Cloud page still presents a Request Access form and the expired “free through August 31, 2026” preview. The newer platform quickstart explicitly describes invite-only beta access, with users without an invitation landing on a Coming Soon page. Please clarify whether that applies to the hosted Cloud API in this PR or to a separate product, and provide the public self-serve signup/onboarding path for the submitted service. A registration form alone does not establish that a new user can obtain usable inference access without an invitation.

2. Public pricing and served limits. Please link an authoritative public rate card matching all four mappings, together with documentation for the served model IDs, 60 RPM / 100,000 TPM limits, and these provider-specific context windows: Qwen 3.5 9B — 65,536; Qwen 3.5 27B — 32,768; Llama 3.2 1B — 24,000; Gemma 4 31B — 8,192. Some model IDs now appear in the public platform documentation, but I could not verify the complete hosted catalog and the numerical prices/limits there. The previously supplied console usage page is not a public rate-card reference. The unauthenticated Models API returns 401, which is expected authentication behavior, not evidence of an outage. Unknown/unpublished RPM or TPM values can be null rather than asserted without a public reference.

3. Canonical API endpoint. providers/openinfer.yaml specifies https://api.openinfer.ai/v1, whereas the current platform quickstart documents https://platform.openinfer.io/v1. Please confirm which endpoint serves the hosted catalog and prices submitted here, and provide the applicable public API documentation. If these are separate services or both URLs remain supported, please explain that distinction rather than changing the URL without verification.

4. Update the PR description. It still describes two models and $0.00 promotional pricing. The actual submission contains four paid provider-model mappings and three new canonical model definitions. Please update the summary, model list, pricing note, and public source links to describe the final submission.

Please do not post or commit an API key; public documentation is sufficient for these questions. Once the sources and access model are clarified and the metadata is aligned, we can re-review the provider addition. Thank you for helping make the entry independently verifiable.