SaladDay · Oct 12, 2026 · Oct 12, 2026
diff --git a/README.md b/README.md
@@ -70,9 +70,11 @@ derives the npm package version from that tag. See [the release guide](docs/rele
 
 ### `mmx text`
 
+`text chat` and `text repl` default to `MiniMax-M3.1-Flash-Preview`. An explicit `--model` or saved `default-text-model` takes precedence. This preview currently requires M Plan or MiniMax Code; with a pay-as-you-go key, select `MiniMax-M3` explicitly. See the [model guide](https://platform.minimax.io/docs/guides/text-generation#agent).
+
 ```bash
 mmx text chat --message "Write a poem"
-mmx text chat --model MiniMax-M3 --message "Hello" --stream
+mmx text chat --model MiniMax-M3.1-Flash-Preview --message "Hello" --stream
 mmx text chat --system "You are a coding assistant" --message "Fizzbuzz in Go"
 mmx text chat --message "user:Hi" --message "assistant:Hey!" --message "How are you?"
 cat messages.json | mmx text chat --messages-file - --output json
@@ -196,22 +198,28 @@ Credential priority: `--api-key` flag > OAuth (config) > `api_key` (config).
 mmx quota
 mmx config show
 mmx config set --key region --value cn
-mmx config set --key default-text-model --value MiniMax-M3
+mmx config set --key default-text-model --value MiniMax-M3.1-Flash-Preview
 mmx config export-schema | jq .
 ```
 
 ### `mmx agent setup`
 
-Configure MiniMax for Claude Code, Codex, Grok Build, OpenCode, Hermes, or Pi while preserving unrelated settings. Run without options for the interactive wizard; supplying options makes the command non-interactive and suitable for scripts.
+Configure MiniMax for Claude Code, Codex, Grok Build, OpenCode, Hermes, Pi, or [Oh my Pi](https://github.com/can1357/oh-my-pi) while preserving unrelated settings. Run without options for the interactive wizard; supplying options makes the command non-interactive and suitable for scripts.
 
 ```bash
 mmx agent setup
 mmx agent setup --agent codex --agent claude-code --api-key "$MINIMAX_API_KEY" --region global
 mmx agent setup --all --api-key "$MINIMAX_API_KEY" --region cn --output json
+mmx agent setup --agent oh-my-pi --api-key "$MINIMAX_API_KEY" --region global
+# --agent omp is an alias for --agent oh-my-pi
 ```
 
+`agent setup` defaults to `MiniMax-M3.1-Flash-Preview` for subscription keys and to `MiniMax-M3` for pay-as-you-go keys (`sk-api-...`). In non-interactive mode, `--model` overrides this selection.
+
 The command verifies the key and selected region before writing. Existing files are backed up when changed; use `--dry-run` to preview without a live request. In the interactive wizard, compatible agents that are missing from `PATH` can be installed from a second multi-select list using their official packages or installer scripts. The Hermes install uses its official core CLI stages and skips optional system packages, Node workspace, browser, computer-use, setup, and gateway stages. After installation, mmx applies the MiniMax configuration while preserving installer settings. Agents that fail a platform or prerequisite check are explained and left configuration-only. Non-interactive invocations remain configuration-only, and the command never launches an agent.
 
+Oh my Pi uses its native `~/.omp/agent/models.yml` and `config.yml` (existing `.yaml` files are updated in place). Configuration honors `OMP_PROFILE` / `PI_PROFILE`, `PI_CONFIG_DIR`, and the default-profile `PI_CODING_AGENT_DIR` override. Run `omp` once to migrate legacy JSON-only configuration before using setup. Interactive installation uses the official prebuilt-binary installer, so it does not require Bun or npm.
+
 ### `mmx update`
 
 ```bash

diff --git a/README_CN.md b/README_CN.md
@@ -65,9 +65,11 @@ mmx quota
 
 ### `mmx text`
 
+`text chat` 和 `text repl` 默认使用 `MiniMax-M3.1-Flash-Preview`；显式指定的 `--model` 或已保存的 `default-text-model` 优先。该预览模型目前仅通过 M Plan 和 MiniMax Code 提供；使用按量付费 Key 时，请显式选择 `MiniMax-M3`。详见[模型调用文档](https://platform.minimax.cn/docs/guides/text-generation#agent)。
+
 ```bash
 mmx text chat --message "写一首诗"
-mmx text chat --model MiniMax-M3 --message "你好" --stream
+mmx text chat --model MiniMax-M3.1-Flash-Preview --message "你好" --stream
 mmx text chat --system "你是编程助手" --message "用 Go 写 Fizzbuzz"
 mmx text chat --message "user:你好" --message "assistant:嗨！" --message "你叫什么名字？"
 cat messages.json | mmx text chat --messages-file - --output json
@@ -170,22 +172,28 @@ Global 与中国两个 region，选用能通过的那个。
 mmx quota
 mmx config show
 mmx config set --key region --value cn
-mmx config set --key default-text-model --value MiniMax-M3
+mmx config set --key default-text-model --value MiniMax-M3.1-Flash-Preview
 mmx config export-schema | jq .
 ```
 
 ### `mmx agent setup`
 
-一键为 Claude Code、Codex、Grok Build、OpenCode、Hermes 或 Pi 配置 MiniMax，并保留配置文件中的其他设置。不带选项时进入交互式向导；带选项时自动使用非交互模式，便于脚本调用。
+一键为 Claude Code、Codex、Grok Build、OpenCode、Hermes、Pi 或 [Oh my Pi](https://github.com/can1357/oh-my-pi) 配置 MiniMax，并保留配置文件中的其他设置。不带选项时进入交互式向导；带选项时自动使用非交互模式，便于脚本调用。
 
 ```bash
 mmx agent setup
 mmx agent setup --agent codex --agent claude-code --api-key "$MINIMAX_API_KEY" --region global
 mmx agent setup --all --api-key "$MINIMAX_API_KEY" --region cn --output json
+mmx agent setup --agent oh-my-pi --api-key "$MINIMAX_API_KEY" --region cn
+# --agent omp 是 --agent oh-my-pi 的别名
 ```
 
+`agent setup` 对订阅 Key 默认使用 `MiniMax-M3.1-Flash-Preview`，对按量付费 Key（`sk-api-...`）默认使用 `MiniMax-M3`。非交互模式下可用 `--model` 覆盖该选择。
+
 写入前会验证 Key 和所选区域；修改已有文件时会创建备份。可用 `--dry-run` 预览且不会发起联网请求。在交互式向导中，兼容且未在 `PATH` 中检测到的 Agent 会出现在第二个安装多选列表，并通过官方软件包或安装脚本安装。Hermes 仅执行官方核心 CLI 安装阶段，跳过可选系统软件包、Node 工作区、浏览器、计算机控制、初始化和网关阶段。安装完成后，mmx 会保留安装器生成的其他设置并写入 MiniMax 配置。平台或依赖检查不通过的 Agent 会说明原因并保持仅配置。非交互调用仍只写配置，且该命令不会启动 Agent。
 
+Oh my Pi 使用原生 `~/.omp/agent/models.yml` 和 `config.yml`（已有 `.yaml` 文件会原地更新），遵循 `OMP_PROFILE` / `PI_PROFILE`、`PI_CONFIG_DIR` 和默认 profile 的 `PI_CODING_AGENT_DIR` 配置。仅有旧版 JSON 配置时，请先运行一次 `omp` 完成迁移。交互式安装使用官方预编译二进制安装器，无需 Bun 或 npm。
+
 ### `mmx update`
 
 ```bash

diff --git a/SDK.md b/SDK.md
@@ -25,16 +25,18 @@ You can also omit `apiKey` if it's already configured via `mmx config set api-ke
 
 ### Text
 
+When `model` is omitted, the SDK uses `MiniMax-M3.1-Flash-Preview`. This preview currently requires M Plan or MiniMax Code; with a pay-as-you-go key, pass `model: 'MiniMax-M3'` explicitly. Thinking is enabled by default with `max` effort; no extra request fields are required. See the [model guide](https://platform.minimax.io/docs/guides/text-generation#agent).
+
 ```typescript
 const response = await sdk.text.chat({
-  model: 'MiniMax-M3',
+  model: 'MiniMax-M3.1-Flash-Preview',
   messages: [{ role: 'user', content: 'Hello!' }],
   max_tokens: 4096,
 });
 
 // Streaming
 const stream = await sdk.text.chat({
-  model: 'MiniMax-M3',
+  model: 'MiniMax-M3.1-Flash-Preview',
   messages: [{ role: 'user', content: 'Write a poem' }],
   stream: true,
 });

diff --git a/skill/SKILL.md b/skill/SKILL.md
@@ -46,7 +46,7 @@ Always use these flags in non-interactive (agent/CI) contexts:
 
 ### text chat
 
-Chat completion. Default model: `MiniMax-M3`.
+Chat completion. Default model: `MiniMax-M3.1-Flash-Preview`. An explicit `--model` or saved `default-text-model` takes precedence. This preview currently requires M Plan or MiniMax Code; with a pay-as-you-go key, use `--model MiniMax-M3`. Thinking is enabled by default with `max` effort. See the [model guide](https://platform.minimax.io/docs/guides/text-generation#agent).
 
 ```bash
 mmx text chat --message <text> [flags]
@@ -57,7 +57,7 @@ mmx text chat --message <text> [flags]
 | `--message <text>` | string, **required**, repeatable | Message text. Prefix with `role:` to set role (e.g. `"system:You are helpful"`, `"user:Hello"`) |
 | `--messages-file <path>` | string | JSON file with messages array. Use `-` for stdin |
 | `--system <text>` | string | System prompt |
-| `--model <model>` | string | Model ID (default: `MiniMax-M3`) |
+| `--model <model>` | string | Model ID (default: `MiniMax-M3.1-Flash-Preview`) |
 | `--max-tokens <n>` | number | Max tokens (default: 4096) |
 | `--temperature <n>` | number | Sampling temperature (0.0, 1.0] |
 | `--top-p <n>` | number | Nucleus sampling threshold |
@@ -381,7 +381,7 @@ Set per-modality defaults so you don't need `--model` every time:
 
 ```bash
 # Set defaults
-mmx config set --key default-text-model --value MiniMax-M3
+mmx config set --key default-text-model --value MiniMax-M3.1-Flash-Preview
 mmx config set --key default-speech-model --value speech-2.8-hd
 mmx config set --key default-video-model --value MiniMax-Hailuo-2.3
 
@@ -391,7 +391,7 @@ mmx speech synthesize --text "Hello" --out hello.mp3
 mmx video generate --prompt "Ocean waves"
 
 # --model still overrides per-call
-mmx text chat --model MiniMax-M3 --message "Hello"
+mmx text chat --model MiniMax-M3.1-Flash-Preview --message "Hello"
 ```
 
 **Resolution priority**: `--model` flag > config default > hardcoded fallback.
diff --git a/src/agent/availability.ts b/src/agent/availability.ts
@@ -10,6 +10,7 @@ const AGENT_EXECUTABLES: Record<AgentId, readonly string[]> = {
   opencode: ['opencode'],
   hermes: ['hermes'],
   pi: ['pi'],
+  'oh-my-pi': ['omp'],
 };
 
 const AGENT_RUNTIME_MARKERS: Partial<Record<AgentId, readonly [string, string]>> = {

diff --git a/src/agent/configurator.ts b/src/agent/configurator.ts
@@ -72,6 +72,45 @@ function pathFromOverride(value: string | undefined, fallback: string, home: str
   return isAbsolute(value) ? value : resolve(home, value);
 }
 
+function ohMyPiConfigPaths(home: string, env: NodeJS.ProcessEnv): string[] {
+  // Match OMP's profile/env precedence, including an explicitly empty OMP_PROFILE.
+  const profile = (env.OMP_PROFILE ?? env.PI_PROFILE)?.trim();
+  const root = join(home, env.PI_CONFIG_DIR || '.omp');
+  const validProfile = (name: string) => /^[a-z0-9][a-z0-9._-]{0,63}$/.test(name)
+    && !name.endsWith('.')
+    && !/^(?:con|prn|aux|nul|com[0-9]|lpt[0-9])(?:\..*)?$/i.test(name);
+  let dir: string;
+  if (profile && profile !== 'default') {
+    if (!validProfile(profile)) {
+      throw new CLIError('Invalid Oh my Pi profile name.', ExitCode.USAGE);
+    }
+    dir = join(root, 'profiles', profile, 'agent');
+  } else {
+    // A parent OMP process propagates its named profile's directory. Selecting
+    // the default profile must discard that inherited value, as OMP itself does.
+    const inheritedProfile = env.PI_PROFILE?.trim();
+    const inheritedDir = inheritedProfile && inheritedProfile !== 'default' && validProfile(inheritedProfile)
+      ? join(root, 'profiles', inheritedProfile, 'agent') : undefined;
+    const override = env.PI_CODING_AGENT_DIR === inheritedDir ? undefined : env.PI_CODING_AGENT_DIR;
+    dir = override ? resolve(override) : join(root, 'agent');
+  }
+  return ['models', 'config'].map(name => {
+    const yml = join(dir, `${name}.yml`);
+    const yaml = join(dir, `${name}.yaml`);
+    if (existsSync(yml)) return yml;
+    if (existsSync(yaml)) return yaml;
+    const legacy = join(dir, name === 'models' ? 'models.json' : 'settings.json');
+    if (existsSync(legacy)) {
+      throw new CLIError(
+        'Oh my Pi has legacy JSON configuration that needs migration.',
+        ExitCode.GENERAL,
+        'Run omp once with the same profile/environment to migrate it to YAML, then retry. No files were changed.',
+      );
+    }
+    return yml;
+  });
+}
+
 function configPathsForAgent(
   agent: AgentId,
   options: AgentSetupOptions,
@@ -115,6 +154,8 @@ function configPathsForAgent(
       );
       return [join(dir, 'models.json'), join(dir, 'settings.json')];
     }
+    case 'oh-my-pi':
+      return ohMyPiConfigPaths(home, env);
   }
 }
 
@@ -938,6 +979,73 @@ function preparePi(options: AgentSetupOptions, paths: string[]): PreparedAgentFi
   ];
 }
 
+function prepareOhMyPi(options: AgentSetupOptions, paths: string[]): PreparedAgentFile[] {
+  const providerId = options.region === 'cn' ? 'minimax-cn' : 'minimax';
+  const baseUrl = endpointsForRegion(options.region).anthropic;
+  return paths.map((path, index) => {
+    const existing = readPrepared(path);
+    const label = `Oh my Pi ${basename(path)}`;
+    const document = parseDocument(existing.before ?? '{}\n');
+    if (document.errors.length > 0) {
+      throw new CLIError(`${label} is not valid YAML.`, ExitCode.GENERAL, INVALID_CONFIG_HINT);
+    }
+    if (document.contents === null) document.set(index === 0 ? 'providers' : 'modelRoles', document.createNode({}));
+    const root = document.toJS() as unknown;
+    assertObject(root, `${label} must contain an object at its root.`);
+    if (index === 0) {
+      if (root.providers !== undefined) assertObject(root.providers, `${label} providers must be an object.`);
+      const provider = root.providers?.[providerId];
+      if (provider !== undefined) assertObject(provider, `${label} provider must be an object.`);
+      const models = provider?.models;
+      if (models !== undefined && !Array.isArray(models)) {
+        throw new CLIError(`${label} models must be an array.`, ExitCode.GENERAL, INVALID_CONFIG_HINT);
+      }
+      const entries = (models ?? []) as unknown[];
+      for (const entry of entries) assertObject(entry, `${label} model entries must be objects.`);
+      const prefix = ['providers', providerId];
+      for (const [key, value] of Object.entries({ baseUrl, apiKey: options.apiKey, api: 'anthropic-messages', auth: 'apiKey' })) {
+        document.setIn([...prefix, key], value);
+      }
+      if (models === undefined) document.setIn([...prefix, 'models'], document.createNode([]));
+      let appendIndex = entries.length;
+      for (const model of MINIMAX_MODELS) {
+        const found = entries.findIndex(entry => (entry as Record<string, unknown>).id === model.id);
+        const modelPath = [...prefix, 'models', found >= 0 ? found : appendIndex++];
+        const definition = {
+          id: model.id,
+          name: model.id,
+          api: 'anthropic-messages',
+          baseUrl,
+          reasoning: true,
+          input: [...model.input],
+          contextWindow: model.contextWindow,
+          maxTokens: model.maxTokens,
+          ...(model.id === 'MiniMax-M3.1-Flash-Preview' ? {
+            thinking: {
+              mode: 'anthropic-adaptive',
+              efforts: ['low', 'medium', 'high', 'xhigh', 'max'],
+              defaultLevel: 'max',
+              requiresEffort: true,
+            },
+          } : {}),
+        };
+        if (found < 0) {
+          document.setIn(modelPath, document.createNode(definition));
+        } else {
+          for (const [key, value] of Object.entries(definition)) {
+            document.setIn([...modelPath, key], document.createNode(value));
+          }
+        }
+      }
+    } else {
+      if (root.modelRoles !== undefined) assertObject(root.modelRoles, `${label} modelRoles must be an object.`);
+      const effort = options.model === 'MiniMax-M3.1-Flash-Preview' ? ':max' : '';
+      document.setIn(['modelRoles', 'default'], `${providerId}/${options.model}${effort}`);
+    }
+    return { agent: 'oh-my-pi', path, ...existing, after: document.toString() };
+  });
+}
+
 export function prepareAgentConfigurations(options: AgentSetupOptions): PreparedAgentFile[] {
   const prepared: PreparedAgentFile[] = [];
   for (const agent of options.agents) {
@@ -961,6 +1069,9 @@ export function prepareAgentConfigurations(options: AgentSetupOptions): Prepared
       case 'pi':
         prepared.push(...preparePi(options, paths));
         break;
+      case 'oh-my-pi':
+        prepared.push(...prepareOhMyPi(options, paths));
+        break;
     }
   }
   assertDistinctConfigurationTargets(prepared);