Skip to content

feat(minimax): upgrade default model to MiniMax-M3 #241

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 6 commits into
base: main
Choose a base branch
from
Open

feat(minimax): upgrade default model to MiniMax-M3 #241

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
267e566
feat(minimax): upgrade default model to MiniMax-M3
Jun 3, 2026
8e27f76
docs: correct MiniMax context window (M3=512K, M2.7=192K; was mistake…
octo-patch Jun 3, 2026
5f1132f
docs: correct MiniMax context window (M3=512K, M2.7=192K; was mistake…
octo-patch Jun 3, 2026
120a1d8
docs: correct MiniMax context window (M3=512K, M2.7=192K; was mistake…
octo-patch Jun 3, 2026
5bf4bd3
docs: correct MiniMax context window (M3=512K, M2.7=192K; was mistake…
octo-patch Jun 3, 2026
6113c73
docs: fix context-window wording (M3=512K, M2.7=192K)
octo-patch Jun 3, 2026
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
93 changes: 41 additions & 52 deletions docs/providers/minimax.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
---
summary: "Use MiniMax M2.5 in OpenClaw"
summary: "Use MiniMax M3 in OpenClaw"
read_when:
- You want MiniMax models in OpenClaw
- You need MiniMax setup guidance
Expand All @@ -8,30 +8,18 @@ title: "MiniMax"

# MiniMax

MiniMax is an AI company that builds the **M2/M2.5** model family. The current
coding-focused release is **MiniMax M2.5** (December 23, 2025), built for
real-world complex tasks.
OpenClaw's MiniMax provider defaults to **MiniMax M3**, the latest generation
reasoning model with a 512K-token context window and image input support.

Source: [MiniMax M2.5 release note](https://www.minimax.io/news/minimax-m25)
Source: [MiniMax](https://www.minimax.io)

## Model overview (M2.5)
## Model overview

MiniMax highlights these improvements in M2.5:
OpenClaw ships the following MiniMax hosted chat models:

- Stronger **multi-language coding** (Rust, Java, Go, C++, Kotlin, Objective-C, TS/JS).
- Better **web/app development** and aesthetic output quality (including native mobile).
- Improved **composite instruction** handling for office-style workflows, building on
interleaved thinking and integrated constraint execution.
- **More concise responses** with lower token usage and faster iteration loops.
- Stronger **tool/agent framework** compatibility and context management (Claude Code,
Droid/Factory AI, Cline, Kilo Code, Roo Code, BlackBox).
- Higher-quality **dialogue and technical writing** outputs.

## MiniMax M2.5 vs MiniMax M2.5 Highspeed

- **Speed:** `MiniMax-M2.5-highspeed` is the official fast tier in MiniMax docs.
- **Cost:** MiniMax pricing lists the same input cost and a higher output cost for highspeed.
- **Current model IDs:** use `MiniMax-M2.5` or `MiniMax-M2.5-highspeed`.
- **`MiniMax-M3`** (default) — latest generation, 512K context, image input.
- **`MiniMax-M2.7`** — previous generation, 200K-token context, text input.
- **`MiniMax-M2.7-highspeed`** — faster M2.7 tier, 200K-token context, text input.
Comment on lines +21 to +22

@coderabbitai coderabbitai Bot Jun 3, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# 描述：搜索 M2.7 模型的上下文窗口定义和使用

# 查找所有上下文窗口相关的常量定义
rg -n -A2 -B2 'CONTEXT_WINDOW|contextWindow' --glob '**/minimax/**' --glob '!node_modules'

# 查找 M2.7 模型定义中的上下文窗口值
ast-grep --pattern $'contextWindow: $_'

Repository: linuxhsj/openclaw-zero-token

Length of output: 11077

🌐 Web query:

MiniMax M2.7 model context window size official documentation

💡 Result:

According to the official MiniMax API documentation, the MiniMax-M2.7 model features a context window size of 204,800 tokens [1][2]. This specification applies to both the standard MiniMax-M2.7 and the high-speed version, MiniMax-M2.7-highspeed [1][3]. Other third-party documentation, such as the NVIDIA NIM reference for this model, also confirms the input context length is 204,800 tokens [4].

Citations:

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# 1) Inspect the referenced doc snippet
echo "---- docs/providers/minimax.md (lines 1-80) ----"
nl -ba docs/providers/minimax.md | sed -n '1,120p'

# 2) Search for any mention of 192K/192k in the repo
echo "---- Searching for 192K/192k ----"
rg -n --hidden --no-ignore-vcs '192[ -]?K|192[ -]?k|192K-token|192k-token|192K' .

# 3) Search for 200K / 200K-token / 200000 in minimax-related docs/code
echo "---- Searching for 200K / 200000 / 204800 ----"
rg -n '200K|200k|200000|204800' docs extensions minimax

# 4) Search for contextWindow assignments specifically for MiniMax-M2.7
echo "---- contextWindow for MiniMax-M2.7 ----"
rg -n --glob '**/*minimax*' 'MiniMax-M2\.7' docs minimax extensions

Repository: linuxhsj/openclaw-zero-token

Length of output: 182

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# 1) Inspect the referenced doc snippet
echo "---- docs/providers/minimax.md (lines 1-120) ----"
nl -ba docs/providers/minimax.md | sed -n '1,120p'

# 2) Search for any mention of 192K/192k in the repo
echo "---- Searching for 192K/192k ----"
rg -n --hidden --no-ignore-vcs '192[ -]?K|192[ -]?k|192K-token|192k-token|192K' .

# 3) Search for 200K / 200000 / 204800 in minimax-related docs/code
echo "---- Searching for 200K / 200000 / 204800 ----"
rg -n '200K|200k|200000|204800' docs extensions minimax

# 4) Search for contextWindow assignments specifically for MiniMax-M2.7
echo "---- contextWindow for MiniMax-M2.7 ----"
rg -n --glob '**/*minimax*' 'MiniMax-M2\.7' docs minimax extensions

Repository: linuxhsj/openclaw-zero-token

Length of output: 183

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# 1) Inspect the referenced doc snippet (no `nl` in sandbox)
echo "---- docs/providers/minimax.md (lines 1-120) ----"
cat -n docs/providers/minimax.md | sed -n '1,140p'

# 2) Search for any mention of 192K/192k in the repo
echo "---- Searching for 192K/192k ----"
rg -n --hidden --no-ignore-vcs '192[ -]?K|192[ -]?k|192K-token|192k-token|192K' .

# 3) Search for 200K / 200000 / 204800 in minimax-related docs/code
echo "---- Searching for 200K / 200000 / 204800 ----"
rg -n '200K|200k|200000|204800' docs extensions minimax

# 4) Search for contextWindow assignments specifically for MiniMax-M2.7
echo "---- contextWindow for MiniMax-M2.7 ----"
rg -n --glob '**/*minimax*' 'MiniMax-M2\.7' docs minimax extensions

Repository: linuxhsj/openclaw-zero-token

Length of output: 4864

🏁 Script executed:

#!/bin/bash
set -euo pipefail

echo "---- Search repo for 192K occurrences (should be none) ----"
rg -n '192[ -]?K|192[ -]?k|192K-token|192k-token|192K' . || true

echo "---- Search docs/providers/minimax.md for 192 ----"
rg -n '192[ -]?K|192[ -]?k|192K-token|192k-token|192K' docs/providers/minimax.md || true

echo "---- Search docs for MiniMax context numbers (200K/200000/204800) ----"
rg -n 'MiniMax-M2\.7|contextWindow:\s*(200000|204800)|200K-token|204800' docs extensions minimax || true

echo "---- Inspect minimax/onboard.test.ts contextWindow section ----"
sed -n '60,110p' minimax/onboard.test.ts || true

echo "---- Inspect extensions/minimax/onboard.test.ts contextWindow section ----"
sed -n '60,110p' extensions/minimax/onboard.test.ts || true

Repository: linuxhsj/openclaw-zero-token

Length of output: 10394

统一 MiniMax-M2.7 上下文窗口文档与代码口径

  • 代码口径：DEFAULT_MINIMAX_CONTEXT_WINDOW = 204800（minimax/model-definitions.ts / minimax/provider-catalog.ts，extensions/minimax 同样为 204800）
  • 文档口径：docs/providers/minimax.md 标注 MiniMax-M2.7 / MiniMax-M2.7-highspeed 为 “200K-token context”，且示例配置使用 contextWindow: 200000；docs/zh-CN/providers/minimax.md 同样为 “20 万 token/200000”
  • 仓库内未检索到 “192K” 相关字样；当前主要不一致是 “200K(200000)” vs “204800”

建议将 docs/providers/minimax.md 与 docs/zh-CN/providers/minimax.md（含示例 JSON5）更新为 204800，或在文档中明确“200K 为四舍五入”的规则，并确保全站一致。

🤖 Prompt for AI Agents 
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/providers/minimax.md` around lines 21 - 22, Docs state a 200000-token
context while code defines DEFAULT_MINIMAX_CONTEXT_WINDOW = 204800 (see
minimax/model-definitions.ts, minimax/provider-catalog.ts and
extensions/minimax); update docs/providers/minimax.md and
docs/zh-CN/providers/minimax.md (including example JSON5) to use 204800 for
MiniMax-M2.7 and MiniMax-M2.7-highspeed, or else add a brief note that “200K” is
a rounded value and show the precise value DEFAULT_MINIMAX_CONTEXT_WINDOW =
204800 for clarity and consistency with the code.


## Choose a setup

Expand All @@ -54,20 +42,20 @@ You will be prompted to select an endpoint:

See [MiniMax OAuth plugin README](https://github.com/openclaw/openclaw/tree/main/extensions/minimax-portal-auth) for details.

### MiniMax M2.5 (API key)
### MiniMax M3 (API key)

**Best for:** hosted MiniMax with Anthropic-compatible API.

Configure via CLI:

- Run `openclaw configure`
- Select **Model/auth**
- Choose **MiniMax M2.5**
- Choose **MiniMax M3**

```json5
{
env: { MINIMAX_API_KEY: "sk-..." },
agents: { defaults: { model: { primary: "minimax/MiniMax-M2.5" } } },
agents: { defaults: { model: { primary: "minimax/MiniMax-M3" } } },
models: {
mode: "merge",
providers: {
Expand All @@ -77,22 +65,22 @@ Configure via CLI:
api: "anthropic-messages",
models: [
{
id: "MiniMax-M2.5",
name: "MiniMax M2.5",
id: "MiniMax-M3",
name: "MiniMax M3",
reasoning: true,
input: ["text"],
cost: { input: 0.3, output: 1.2, cacheRead: 0.03, cacheWrite: 0.12 },
contextWindow: 200000,
maxTokens: 8192,
input: ["text", "image"],
cost: { input: 0.6, output: 2.4, cacheRead: 0.12, cacheWrite: 0 },
contextWindow: 512_000,
maxTokens: 131072,
},
{
id: "MiniMax-M2.5-highspeed",
name: "MiniMax M2.5 Highspeed",
id: "MiniMax-M2.7",
name: "MiniMax M2.7",
reasoning: true,
input: ["text"],
cost: { input: 0.3, output: 1.2, cacheRead: 0.03, cacheWrite: 0.12 },
cost: { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0.375 },
contextWindow: 200000,
maxTokens: 8192,
maxTokens: 131072,
},
],
},
Expand All @@ -101,9 +89,9 @@ Configure via CLI:
}
```

### MiniMax M2.5 as fallback (example)
### MiniMax M3 as fallback (example)

**Best for:** keep your strongest latest-generation model as primary, fail over to MiniMax M2.5.
**Best for:** keep your strongest latest-generation model as primary, fail over to MiniMax M3.
Example below uses Opus as a concrete primary; swap to your preferred latest-gen primary model.

```json5
Expand All @@ -113,11 +101,11 @@ Example below uses Opus as a concrete primary; swap to your preferred latest-gen
defaults: {
models: {
"anthropic/claude-opus-4-6": { alias: "primary" },
"minimax/MiniMax-M2.5": { alias: "minimax" },
"minimax/MiniMax-M3": { alias: "minimax" },
},
model: {
primary: "anthropic/claude-opus-4-6",
fallbacks: ["minimax/MiniMax-M2.5"],
fallbacks: ["minimax/MiniMax-M3"],
},
},
},
Expand All @@ -127,7 +115,7 @@ Example below uses Opus as a concrete primary; swap to your preferred latest-gen
### Optional: Local via LM Studio (manual)

**Best for:** local inference with LM Studio.
We have seen strong results with MiniMax M2.5 on powerful hardware (e.g. a
We have seen strong results with MiniMax on powerful hardware (e.g. a
desktop/server) using LM Studio's local server.

Configure manually via `openclaw.json`:
Expand All @@ -136,8 +124,8 @@ Configure manually via `openclaw.json`:
{
agents: {
defaults: {
model: { primary: "lmstudio/minimax-m2.5-gs32" },
models: { "lmstudio/minimax-m2.5-gs32": { alias: "Minimax" } },
model: { primary: "lmstudio/minimax-m3-gs32" },
models: { "lmstudio/minimax-m3-gs32": { alias: "Minimax" } },
},
},
models: {
Expand All @@ -149,8 +137,8 @@ Configure manually via `openclaw.json`:
api: "openai-responses",
models: [
{
id: "minimax-m2.5-gs32",
name: "MiniMax M2.5 GS32",
id: "minimax-m3-gs32",
name: "MiniMax M3 GS32",
reasoning: false,
input: ["text"],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
Expand All @@ -170,7 +158,7 @@ Use the interactive config wizard to set MiniMax without editing JSON:

1. Run `openclaw configure`.
2. Select **Model/auth**.
3. Choose **MiniMax M2.5**.
3. Choose **MiniMax M3**.
4. Pick your default model when prompted.

## Configuration options
Expand All @@ -185,30 +173,31 @@ Use the interactive config wizard to set MiniMax without editing JSON:
## Notes

- Model refs are `minimax/<model>`.
- Recommended model IDs: `MiniMax-M2.5` and `MiniMax-M2.5-highspeed`.
- Recommended model IDs: `MiniMax-M3` (default), `MiniMax-M2.7`, and `MiniMax-M2.7-highspeed`.
- Coding Plan usage API: `https://api.minimaxi.com/v1/api/openplatform/coding_plan/remains` (requires a coding plan key).
- Update pricing values in `models.json` if you need exact cost tracking.
- Referral link for MiniMax Coding Plan (10% off): [https://platform.minimax.io/subscribe/coding-plan?code=DbXJTRClnb&source=link](https://platform.minimax.io/subscribe/coding-plan?code=DbXJTRClnb&source=link)
- See [/concepts/model-providers](/concepts/model-providers) for provider rules.
- Use `openclaw models list` and `openclaw models set minimax/MiniMax-M2.5` to switch.
- Use `openclaw models list` and `openclaw models set minimax/MiniMax-M3` to switch.

## Troubleshooting

### “Unknown model: minimax/MiniMax-M2.5”
### "Unknown model: minimax/MiniMax-M3"

This usually means the **MiniMax provider isn’t configured** (no provider entry
This usually means the **MiniMax provider isn't configured** (no provider entry
and no MiniMax auth profile/env key found). A fix for this detection is in
**2026.1.12** (unreleased at the time of writing). Fix by:

- Upgrading to **2026.1.12** (or run from source `main`), then restarting the gateway.
- Running `openclaw configure` and selecting **MiniMax M2.5**, or
- Running `openclaw configure` and selecting **MiniMax M3**, or
- Adding the `models.providers.minimax` block manually, or
- Setting `MINIMAX_API_KEY` (or a MiniMax auth profile) so the provider can be injected.

Make sure the model id is **case‑sensitive**:
Make sure the model id is **case-sensitive**:

- `minimax/MiniMax-M2.5`
- `minimax/MiniMax-M2.5-highspeed`
- `minimax/MiniMax-M3`
- `minimax/MiniMax-M2.7`
- `minimax/MiniMax-M2.7-highspeed`

Then recheck with:

Expand Down
81 changes: 41 additions & 40 deletions docs/zh-CN/providers/minimax.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
read_when:
- 你想在 OpenClaw 中使用 MiniMax 模型
- 你需要 MiniMax 设置指南
summary: 在 OpenClaw 中使用 MiniMax M2.1
summary: 在 OpenClaw 中使用 MiniMax M3

@coderabbitai coderabbitai Bot Jun 3, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟠 Major | ⚡ Quick win

请改为走 zh-CN 文档生成流水线，不要直接手改生成文件。

本次在 docs/zh-CN/providers/minimax.md 的直接内容修改会引入与 x-i18n.source_hash 的漂移风险，后续执行 i18n 同步时可能被覆盖。请改为：先更新英文文档，再按仓库流程更新术语并运行 scripts/docs-i18n，仅在被明确要求时做定点修订。

As per coding guidelines, "docs/zh-CN/** is generated; do not edit unless the user explicitly asks. Pipeline: update English docs → adjust glossary (docs/.i18n/glossary.zh-CN.json) → run scripts/docs-i18n → apply targeted fixes only if instructed".

Also applies to: 18-20, 51-51, 59-59, 64-90, 98-100, 109-114, 123-123, 131-132, 144-145, 165-165, 184-184, 188-188, 193-193, 199-201

🤖 Prompt for AI Agents 
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/zh-CN/providers/minimax.md` at line 5, Summary: You edited a generated
file (docs/zh-CN/providers/minimax.md) directly; generated docs must not be
hand-modified. Fix: revert changes to docs/zh-CN/providers/minimax.md, then make
your authoritative changes in the English source, update the Chinese glossary at
docs/.i18n/glossary.zh-CN.json as needed, and regenerate translations by running
scripts/docs-i18n; ensure x-i18n.source_hash remains consistent and only perform
targeted edits to docs/zh-CN/** if explicitly requested by the i18n process.

title: MiniMax
x-i18n:
generated_at: "2026-02-03T10:08:52Z"
Expand All @@ -15,26 +15,17 @@ x-i18n:

# MiniMax

MiniMax 是一家构建 **M2/M2.1** 模型系列的 AI 公司。当前面向编程的版本是 **MiniMax M2.1**（2025 年 12 月 23 日），专为现实世界的复杂任务而构建。
OpenClaw 的 MiniMax 提供商默认使用 **MiniMax M3**，这是最新一代的推理模型，拥有 100 万 token 上下文窗口和图片输入支持。

来源：[MiniMax M2.1 发布说明](https://www.minimax.io/news/minimax-m21)
来源：[MiniMax](https://www.minimax.io)

## 模型概述（M2.1）
## 模型概述

MiniMax 强调 M2.1 的以下改进：
OpenClaw 内置以下 MiniMax 托管的聊天模型：

- 更强的**多语言编程**能力（Rust、Java、Go、C++、Kotlin、Objective-C、TS/JS）。
- 更好的 **Web/应用开发**和美观输出质量（包括原生移动端）。
- 改进的**复合指令**处理，适用于办公风格的工作流程，基于交错思考和集成约束执行。
- **更简洁的响应**，更低的 token 使用量和更快的迭代循环。
- 更强的**工具/智能体框架**兼容性和上下文管理（Claude Code、Droid/Factory AI、Cline、Kilo Code、Roo Code、BlackBox）。
- 更高质量的**对话和技术写作**输出。

## MiniMax M2.1 vs MiniMax M2.1 Lightning

- **速度：** Lightning 是 MiniMax 定价文档中的"快速"变体。
- **成本：** 定价显示相同的输入成本，但 Lightning 的输出成本更高。
- **编程计划路由：** Lightning 后端在 MiniMax 编程计划中不能直接使用。MiniMax 自动将大多数请求路由到 Lightning，但在流量高峰期会回退到常规 M2.1 后端。
- **`MiniMax-M3`**（默认）— 最新一代，100 万 token 上下文，支持图片输入。
- **`MiniMax-M2.7`** — 上一代，20 万 token 上下文，仅文本输入。
- **`MiniMax-M2.7-highspeed`** — M2.7 的更快速版本，20 万 token 上下文，仅文本输入。

## 选择设置方式

Expand All @@ -57,20 +48,20 @@ openclaw onboard --auth-choice minimax-portal

详情参见 [MiniMax OAuth 插件 README](https://github.com/openclaw/openclaw/tree/main/extensions/minimax-portal-auth)。

### MiniMax M2.1（API 密钥）
### MiniMax M3（API 密钥）

**适用于：** 使用 Anthropic 兼容 API 的托管 MiniMax。

通过 CLI 配置：

- 运行 `openclaw configure`
- 选择 **Model/auth**
- 选择 **MiniMax M2.1**
- 选择 **MiniMax M3**

```json5
{
env: { MINIMAX_API_KEY: "sk-..." },
agents: { defaults: { model: { primary: "minimax/MiniMax-M2.1" } } },
agents: { defaults: { model: { primary: "minimax/MiniMax-M3" } } },
models: {
mode: "merge",
providers: {
Expand All @@ -80,13 +71,22 @@ openclaw onboard --auth-choice minimax-portal
api: "anthropic-messages",
models: [
{
id: "MiniMax-M2.1",
name: "MiniMax M2.1",
reasoning: false,
id: "MiniMax-M3",
name: "MiniMax M3",
reasoning: true,
input: ["text", "image"],
cost: { input: 0.6, output: 2.4, cacheRead: 0.12, cacheWrite: 0 },
contextWindow: 512_000,
maxTokens: 131072,
},
{
id: "MiniMax-M2.7",
name: "MiniMax M2.7",
reasoning: true,
input: ["text"],
cost: { input: 15, output: 60, cacheRead: 2, cacheWrite: 10 },
cost: { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0.375 },
contextWindow: 200000,
maxTokens: 8192,
maxTokens: 131072,
},
],
},
Expand All @@ -95,9 +95,9 @@ openclaw onboard --auth-choice minimax-portal
}
```

### MiniMax M2.1 作为备用（Opus 为主）
### MiniMax M3 作为备用（Opus 为主）

**适用于：** 保持 Opus 4.5 为主模型，故障时切换到 MiniMax M2.1。
**适用于：** 保持 Opus 4.5 为主模型，故障时切换到 MiniMax M3。

```json5
{
Expand All @@ -106,11 +106,11 @@ openclaw onboard --auth-choice minimax-portal
defaults: {
models: {
"anthropic/claude-opus-4-5": { alias: "opus" },
"minimax/MiniMax-M2.1": { alias: "minimax" },
"minimax/MiniMax-M3": { alias: "minimax" },
},
model: {
primary: "anthropic/claude-opus-4-5",
fallbacks: ["minimax/MiniMax-M2.1"],
fallbacks: ["minimax/MiniMax-M3"],
},
},
},
Expand All @@ -120,16 +120,16 @@ openclaw onboard --auth-choice minimax-portal
### 可选：通过 LM Studio 本地运行（手动）

**适用于：** 使用 LM Studio 进行本地推理。
我们在强大硬件（例如台式机/服务器）上使用 LM Studio 的本地服务器运行 MiniMax M2.1 时看到了出色的效果。
我们在强大硬件（例如台式机/服务器）上使用 LM Studio 的本地服务器运行 MiniMax 时看到了出色的效果。

通过 `openclaw.json` 手动配置：

```json5
{
agents: {
defaults: {
model: { primary: "lmstudio/minimax-m2.1-gs32" },
models: { "lmstudio/minimax-m2.1-gs32": { alias: "Minimax" } },
model: { primary: "lmstudio/minimax-m3-gs32" },
models: { "lmstudio/minimax-m3-gs32": { alias: "Minimax" } },
},
},
models: {
Expand All @@ -141,8 +141,8 @@ openclaw onboard --auth-choice minimax-portal
api: "openai-responses",
models: [
{
id: "minimax-m2.1-gs32",
name: "MiniMax M2.1 GS32",
id: "minimax-m3-gs32",
name: "MiniMax M3 GS32",
reasoning: false,
input: ["text"],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
Expand All @@ -162,7 +162,7 @@ openclaw onboard --auth-choice minimax-portal

1. 运行 `openclaw configure`。
2. 选择 **Model/auth**。
3. 选择 **MiniMax M2.1**。
3. 选择 **MiniMax M3**。
4. 在提示时选择你的默认模型。

## 配置选项
Expand All @@ -181,23 +181,24 @@ openclaw onboard --auth-choice minimax-portal
- 如果需要精确的成本跟踪，请更新 `models.json` 中的定价值。
- MiniMax 编程计划推荐链接（9 折优惠）：https://platform.minimax.io/subscribe/coding-plan?code=DbXJTRClnb&source=link
- 参见 [/concepts/model-providers](/concepts/model-providers) 了解提供商规则。
- 使用 `openclaw models list` 和 `openclaw models set minimax/MiniMax-M2.1` 切换模型。
- 使用 `openclaw models list` 和 `openclaw models set minimax/MiniMax-M3` 切换模型。

## 故障排除

### "Unknown model: minimax/MiniMax-M2.1"
### "Unknown model: minimax/MiniMax-M3"

这通常意味着 **MiniMax 提供商未配置**（没有提供商条目，也没有找到 MiniMax 认证配置文件/环境变量密钥）。此检测的修复在 **2026.1.12** 中（撰写本文时尚未发布）。修复方法：

- 升级到 **2026.1.12**（或从源码 `main` 分支运行），然后重启 Gateway 网关。
- 运行 `openclaw configure` 并选择 **MiniMax M2.1**，或
- 运行 `openclaw configure` 并选择 **MiniMax M3**，或
- 手动添加 `models.providers.minimax` 块，或
- 设置 `MINIMAX_API_KEY`（或 MiniMax 认证配置文件）以便注入提供商。

确保模型 id **区分大小写**：

- `minimax/MiniMax-M2.1`
- `minimax/MiniMax-M2.1-lightning`
- `minimax/MiniMax-M3`
- `minimax/MiniMax-M2.7`
- `minimax/MiniMax-M2.7-highspeed`

然后重新检查：

Expand Down
Loading
Loading