Skip to content

feat: upgrade MiniMax default model to M3 #13

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
octo-patch wants to merge 1 commit into
base: main
Choose a base branch
from
Open

feat: upgrade MiniMax default model to M3 #13

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .env.example
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@ LLM_API_KEY=
# Local or Remote LLM URL (Ollama example: http://127.0.0.1:11434/v1)
LLM_BASE_URL=

# Model Name (e.g., gpt-4o-mini, MiniMax-M2.5, deepseek-chat, llama3)
# Model Name (e.g., gpt-4o-mini, MiniMax-M3, deepseek-chat, llama3)
LLM_MODEL=

# Provider-specific API keys (used when LLM_API_KEY is not set)
Expand Down
4 changes: 2 additions & 2 deletions README-zh.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -163,7 +163,7 @@ cp .env.example .env
- *💡 零成本本地提示：如果你在使用本地部署大模型（如 Ollama + llama3），可以将其设为 `local-fallback`。*
- **`LLM_PROVIDER=`**：选择内置供应商预设（`openai`、`minimax`、`deepseek`、`ollama`）。省略时会根据 `LLM_BASE_URL` 或对应的 API Key 环境变量自动检测。
- **`LLM_BASE_URL=`**：API转发地址（例如：`https://api.openai.com/v1` 或 本地 `http://127.0.0.1:11434/v1`）。
- **`LLM_MODEL=`**：要执行推理的模型名字（如 `gpt-4o-mini`、`MiniMax-M2.5`）。
- **`LLM_MODEL=`**：要执行推理的模型名字（如 `gpt-4o-mini`、`MiniMax-M3`）。
- **`DISCOVER_BATCH_SIZE`** / **`EVALUATE_BATCH_SIZE`** / **`UPDATE_STATUS_BATCH_SIZE`**：控制每次探索拉取的个数、一次批量合并扔给 AI 判断的项目个数，以及状态更新时的批处理大小。
- **`LOOP_INTERVAL_SECONDS`**: 可调整 `ai:loop-eval` 循环模式每次休息的打底时间（默认 60 秒）。
- **`UPDATE_STATUS_INTERVAL_SECONDS`**: 可调整 `ai:update-status:loop` 状态更新循环模式每次休息的打底时间（默认 60 秒）。
Expand All @@ -180,7 +180,7 @@ cp .env.example .env
| 供应商 | `LLM_PROVIDER` | 默认模型 | API Key 环境变量 |
|--------|----------------|----------|-----------------|
| [OpenAI](https://openai.com) | `openai` | `gpt-4o-mini` | `OPENAI_API_KEY` 或 `LLM_API_KEY` |
| [MiniMax](https://www.minimaxi.com) | `minimax` | `MiniMax-M2.5` | `MINIMAX_API_KEY` 或 `LLM_API_KEY` |
| [MiniMax](https://www.minimaxi.com) | `minimax` | `MiniMax-M3` | `MINIMAX_API_KEY` 或 `LLM_API_KEY` |
| [DeepSeek](https://deepseek.com) | `deepseek` | `deepseek-chat` | `DEEPSEEK_API_KEY` 或 `LLM_API_KEY` |
| [Ollama](https://ollama.ai) (本地) | `ollama` | `llama3` | 不需要（使用 `local-fallback`） |

Expand Down
4 changes: 2 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -140,7 +140,7 @@ Open `.env` and adjust the core configurations:
- *💡 Zero-Cost Prompt: If you are using a local LLM setup (e.g. Ollama via llama3), you can simply use `LLM_API_KEY=local-fallback`.*
- **`LLM_PROVIDER=`**: Select a built-in provider preset (`openai`, `minimax`, `deepseek`, `ollama`). When omitted, auto-detected from `LLM_BASE_URL` or provider-specific API key env vars.
- **`LLM_BASE_URL=`**: LLM endpoint (e.g. `https://api.openai.com/v1`, or local `http://127.0.0.1:11434/v1`).
- **`LLM_MODEL=`**: Standard model identity to use (e.g. `gpt-4o-mini`, `MiniMax-M2.5`).
- **`LLM_MODEL=`**: Standard model identity to use (e.g. `gpt-4o-mini`, `MiniMax-M3`).
- **`DISCOVER_BATCH_SIZE`** / **`EVALUATE_BATCH_SIZE`** / **`UPDATE_STATUS_BATCH_SIZE`**: Modify limits per pull from GitHub, per LLM prompt, and for status update batching.
- **`LOOP_INTERVAL_SECONDS`**: Configure the base idle time interval between consecutive `ai:loop-eval` cycles (default: 60s).
- **`MAX_PAGES_DEFAULT`**: Default max pages to explore per topic (default: 5).
Expand All @@ -156,7 +156,7 @@ The evaluation engine supports any **OpenAI-compatible** LLM API. Built-in prese
| Provider | `LLM_PROVIDER` | Default Model | API Key Env |
|----------|----------------|---------------|-------------|
| [OpenAI](https://openai.com) | `openai` | `gpt-4o-mini` | `OPENAI_API_KEY` or `LLM_API_KEY` |
| [MiniMax](https://www.minimaxi.com) | `minimax` | `MiniMax-M2.5` | `MINIMAX_API_KEY` or `LLM_API_KEY` |
| [MiniMax](https://www.minimaxi.com) | `minimax` | `MiniMax-M3` | `MINIMAX_API_KEY` or `LLM_API_KEY` |
| [DeepSeek](https://deepseek.com) | `deepseek` | `deepseek-chat` | `DEEPSEEK_API_KEY` or `LLM_API_KEY` |
| [Ollama](https://ollama.ai) (local) | `ollama` | `llama3` | N/A (uses `local-fallback`) |

Expand Down
6 changes: 3 additions & 3 deletions scripts/llm-provider.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,10 +18,10 @@ const PROVIDER_PRESETS = {
minimax: {
name: 'MiniMax',
baseUrl: 'https://api.minimax.io/v1',
defaultModel: 'MiniMax-M2.5',
defaultModel: 'MiniMax-M3',
envKey: 'MINIMAX_API_KEY',
temperatureRange: [0, 1],
models: ['MiniMax-M2.7', 'MiniMax-M2.5', 'MiniMax-M2.5-highspeed'],
models: ['MiniMax-M3', 'MiniMax-M2.7'],
},
deepseek: {
name: 'DeepSeek',
Expand Down Expand Up @@ -132,7 +132,7 @@ function buildRequestBody(provider, model, messages, options = {}) {
}

/**
* Strip <think>…</think> tags that some models (e.g. MiniMax-M2.5) may
* Strip <think>…</think> tags that some models (e.g. MiniMax-M3) may
* include in their responses.
*/
function stripThinkTags(text) {
Expand Down
6 changes: 3 additions & 3 deletions tests/llm-provider.integration.test.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ describe('Integration: provider config resolution', () => {
responseFormat: { type: 'json_object' },
});

expect(body.model).toBe('MiniMax-M2.5');
expect(body.model).toBe('MiniMax-M3');
expect(body.temperature).toBe(0.1);
expect(body.response_format).toEqual({ type: 'json_object' });
expect(body.messages).toHaveLength(2);
Expand All @@ -59,7 +59,7 @@ describe('Integration: provider config resolution', () => {
const config = resolveLLMConfig();
expect(config.provider).toBe('minimax');
expect(config.baseUrl).toBe('https://api.minimax.io/v1');
expect(config.model).toBe('MiniMax-M2.5');
expect(config.model).toBe('MiniMax-M3');
expect(config.apiKey).toBe('mm-test-key');
});

Expand Down Expand Up @@ -101,7 +101,7 @@ describe('Integration: live MiniMax API call', () => {
provider: 'minimax',
baseUrl: 'https://api.minimax.io/v1',
apiKey,
model: 'MiniMax-M2.5-highspeed',
model: 'MiniMax-M3',
};

const messages = [
Expand Down
21 changes: 10 additions & 11 deletions tests/llm-provider.test.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,18 +23,17 @@ describe('PROVIDER_PRESETS', () => {
expect(PROVIDER_PRESETS.minimax.baseUrl).toBe('https://api.minimax.io/v1');
});

it('minimax preset should default to MiniMax-M2.5 model', () => {
expect(PROVIDER_PRESETS.minimax.defaultModel).toBe('MiniMax-M2.5');
it('minimax preset should default to MiniMax-M3 model', () => {
expect(PROVIDER_PRESETS.minimax.defaultModel).toBe('MiniMax-M3');
});

it('minimax preset should have temperature range [0, 1]', () => {
expect(PROVIDER_PRESETS.minimax.temperatureRange).toEqual([0, 1]);
});

it('minimax preset should list available models', () => {
expect(PROVIDER_PRESETS.minimax.models).toContain('MiniMax-M3');
expect(PROVIDER_PRESETS.minimax.models).toContain('MiniMax-M2.7');
expect(PROVIDER_PRESETS.minimax.models).toContain('MiniMax-M2.5');
expect(PROVIDER_PRESETS.minimax.models).toContain('MiniMax-M2.5-highspeed');
});

it('minimax preset should use MINIMAX_API_KEY env var', () => {
Expand Down Expand Up @@ -157,7 +156,7 @@ describe('resolveLLMConfig', () => {
const config = resolveLLMConfig();
expect(config.provider).toBe('minimax');
expect(config.baseUrl).toBe('https://api.minimax.io/v1');
expect(config.model).toBe('MiniMax-M2.5');
expect(config.model).toBe('MiniMax-M3');
});

it('should use provider-specific env key when LLM_API_KEY is empty', () => {
Expand Down Expand Up @@ -201,7 +200,7 @@ describe('resolveLLMConfig', () => {
/* ------------------------------------------------------------------ */
describe('buildRequestBody', () => {
it('should clamp temperature for minimax provider', () => {
const body = buildRequestBody('minimax', 'MiniMax-M2.5', [], {
const body = buildRequestBody('minimax', 'MiniMax-M3', [], {
temperature: 1.5,
});
expect(body.temperature).toBe(1);
Expand All @@ -215,26 +214,26 @@ describe('buildRequestBody', () => {
});

it('should include response_format when provided', () => {
const body = buildRequestBody('minimax', 'MiniMax-M2.5', [], {
const body = buildRequestBody('minimax', 'MiniMax-M3', [], {
responseFormat: { type: 'json_object' },
});
expect(body.response_format).toEqual({ type: 'json_object' });
});

it('should not include response_format when not provided', () => {
const body = buildRequestBody('minimax', 'MiniMax-M2.5', []);
const body = buildRequestBody('minimax', 'MiniMax-M3', []);
expect(body).not.toHaveProperty('response_format');
});

it('should set model and messages correctly', () => {
const msgs = [{ role: 'user', content: 'Hello' }];
const body = buildRequestBody('minimax', 'MiniMax-M2.5', msgs);
expect(body.model).toBe('MiniMax-M2.5');
const body = buildRequestBody('minimax', 'MiniMax-M3', msgs);
expect(body.model).toBe('MiniMax-M3');
expect(body.messages).toBe(msgs);
});

it('should default temperature to 0.1 when not specified', () => {
const body = buildRequestBody('minimax', 'MiniMax-M2.5', []);
const body = buildRequestBody('minimax', 'MiniMax-M3', []);
expect(body.temperature).toBe(0.1);
});
});
Expand Down