---
title: MiniMax M3 | Runware Docs
url: https://runware.ai/docs/models/minimax-m3
description: Frontier multimodal agentic model with 1M context for coding, long-horizon tasks, and visual understanding
---
# MiniMax M3

MiniMax M3 is MiniMax's new flagship open-weight model for coding, agentic execution, and multimodal reasoning. It is built around the MiniMax Sparse Attention architecture, supports up to 1 million tokens of context, accepts image and video input in addition to text, and is designed for long-running software, research, browsing, and desktop-operation workflows that need strong tool use and sustained multi-step performance.

- **ID**: `minimax-m3`
- **Status**: coming-soon
- **Creator**: MiniMax
- **Release Date**: June 1, 2026
- **Capabilities**: Text to Text, Image to Text, Video to Text

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/minimax-m3#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `textInference`

Identifier for the type of task being performed

### [taskUUID](https://runware.ai/docs/models/minimax-m3#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [outputFormat](https://runware.ai/docs/models/minimax-m3#request-outputformat)

- **Type**: `string`
- **Default**: `TEXT`

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

**Allowed values**: `TEXT`

### [webhookURL](https://runware.ai/docs/models/minimax-m3#request-webhookurl)

- **Type**: `string`
- **Format**: `URI`

Specifies a webhook URL where JSON responses will be sent via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/platform/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/minimax-m3#request-deliverymethod)

- **Type**: `string`
- **Default**: `sync`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns complete results directly in the API response.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.
- `stream` Streams results token-by-token as they are generated.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/platform/task-polling) (platform)

### [includeCost](https://runware.ai/docs/models/minimax-m3#request-includecost)

- **Type**: `boolean`
- **Default**: `false`

Include task cost in the response.

### [includeUsage](https://runware.ai/docs/models/minimax-m3#request-includeusage)

- **Type**: `boolean`
- **Default**: `false`

Include token usage statistics in the response.

### [numberResults](https://runware.ai/docs/models/minimax-m3#request-numberresults)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `4`
- **Default**: `1`

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/minimax-m3#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `minimax-m3`

Identifier of the model to use for generation.

### [messages](https://runware.ai/docs/models/minimax-m3#request-messages)

- **Path**: `messages.role`
- **Type**: `array of objects (2 properties)`
- **Required**: true

Array of chat messages forming the conversation context.

#### [role](https://runware.ai/docs/models/minimax-m3#request-messages-role)

- **Path**: `messages.role`
- **Type**: `string`
- **Required**: true

The role of the message author.

**Allowed values**: `user` `assistant`

#### [content](https://runware.ai/docs/models/minimax-m3#request-messages-content)

- **Path**: `messages.content`
- **Type**: `string`
- **Required**: true
- **Min**: `1`

The text content of the message.

**Settings**

Technical parameters to fine-tune the inference process. These must be nested inside the \`settings\` object.

### [systemPrompt](https://runware.ai/docs/models/minimax-m3#request-settings-systemprompt)

- **Path**: `settings.systemPrompt`
- **Type**: `string`
- **Min**: `1`
- **Max**: `200000`

System-level instruction that guides the model's behavior and output style across the entire generation.

### [temperature](https://runware.ai/docs/models/minimax-m3#request-settings-temperature)

- **Path**: `settings.temperature`
- **Type**: `float`
- **Min**: `0`
- **Max**: `2`
- **Step**: `0.01`
- **Default**: `1`

Controls randomness in generation. Lower values produce more deterministic outputs, higher values increase variation and creativity.

### [topP](https://runware.ai/docs/models/minimax-m3#request-settings-topp)

- **Path**: `settings.topP`
- **Type**: `float`
- **Min**: `0`
- **Max**: `1`
- **Step**: `0.01`
- **Default**: `0.95`

Nucleus sampling parameter that controls diversity by limiting the probability mass. Lower values make outputs more focused, higher values increase diversity.

### [maxTokens](https://runware.ai/docs/models/minimax-m3#request-settings-maxtokens)

- **Path**: `settings.maxTokens`
- **Type**: `integer`
- **Min**: `1`
- **Max**: `512000`
- **Default**: `128000`

Maximum number of tokens to generate in the response.

### [toolChoice](https://runware.ai/docs/models/minimax-m3#request-toolchoice)

- **Path**: `toolChoice.type`
- **Type**: `object (2 properties)`

Controls how the model selects which tool to call. This only takes effect when `tools` are defined.

**Examples**:

**Let the model decide (default):**

```json
"toolChoice": {
  "type": "auto"
}
```

**Force a specific tool call:**

```json
"toolChoice": {
  "type": "tool",
  "name": "get_weather"
}
```

**Require any tool call:**

```json
"toolChoice": {
  "type": "any"
}
```

#### [type](https://runware.ai/docs/models/minimax-m3#request-toolchoice-type)

- **Path**: `toolChoice.type`
- **Type**: `string`
- **Required**: true

Strategy the model uses to decide when and which tools to call.

**Allowed values**:

- `auto` The model decides whether to call a tool based on the conversation context. This is the recommended default.
- `any` The model must call at least one tool but chooses which one. Useful when you always need structured output.
- `tool` The model must call the specific tool identified by name. Use this to force a particular function call.
- `none` The model will not call any tool, even if tools are defined. Useful for forcing a text-only response.

#### [name](https://runware.ai/docs/models/minimax-m3#request-toolchoice-name)

- **Path**: `toolChoice.name`
- **Type**: `string`

Name of the specific tool the model must call. Required when type is `tool`.

### [tools](https://runware.ai/docs/models/minimax-m3#request-tools)

- **Path**: `tools.type`
- **Type**: `array of objects (4 properties)`

An array of tool definitions that the model may call during generation. The model can invoke one or more tools based on the conversation context, outputting structured calls with arguments instead of (or alongside) free-text.

For `function` tools, each definition requires:

- **`type`**: `"function"`
- **`name`**: Unique identifier (alphanumeric, hyphens, underscores; max 64 chars).
- **`description`**: What the function does. The model uses this to decide when to call it.
- **`schema`**: JSON Schema object describing the expected input arguments.

> [!NOTE]
> The `search` tool is executed server-side by the provider. You don't need to handle the tool result yourself.

> [!NOTE]
> The `codeInterpreter` tool is executed server-side by the provider. You don't need to handle the tool result yourself.

**Examples**:

**Function tool, weather lookup:**

```json
"tools": [
  {
    "type": "function",
    "name": "get_weather",
    "description": "Get current weather for a city",
    "schema": {
      "type": "object",
      "properties": {
        "city": { "type": "string", "description": "City name" }
      },
      "required": ["city"]
    }
  }
],
"toolChoice": { "type": "auto" }
```

**Built-in web search:**

```json
"tools": [
  { "type": "search" }
]
```

**Built-in code interpreter:**

```json
"tools": [
  { "type": "codeInterpreter" }
]
```

**Multiple function tools:**

```json
"tools": [
  {
    "type": "function",
    "name": "search_products",
    "description": "Search the product catalog by query and filters.",
    "schema": {
      "type": "object",
      "properties": {
        "query": { "type": "string" },
        "category": { "type": "string" }
      },
      "required": ["query"]
    }
  },
  {
    "type": "function",
    "name": "add_to_cart",
    "description": "Add a product to the user's shopping cart.",
    "schema": {
      "type": "object",
      "properties": {
        "productId": { "type": "string" },
        "quantity": { "type": "integer", "minimum": 1 }
      },
      "required": ["productId"]
    }
  }
]
```

#### [type](https://runware.ai/docs/models/minimax-m3#request-tools-type)

- **Path**: `tools.type`
- **Type**: `string`
- **Required**: true

The kind of tool to make available to the model. User-defined functions require `name` and `schema`, while built-in tools (`search`, `codeInterpreter`) are executed server-side by the provider.

**Allowed values**: `function`

#### [name](https://runware.ai/docs/models/minimax-m3#request-tools-name)

- **Path**: `tools.name`
- **Type**: `string`
- **Max**: `64`

Unique function name. Required for function tools.

#### [description](https://runware.ai/docs/models/minimax-m3#request-tools-description)

- **Path**: `tools.description`
- **Type**: `string`

Explanation of what the function does, used by the model to decide when to call it.

#### [schema](https://runware.ai/docs/models/minimax-m3#request-tools-schema)

- **Path**: `tools.schema`
- **Type**: `object`

JSON Schema object describing the function's input parameters.

## Response Parameters

### [taskType](https://runware.ai/docs/models/minimax-m3#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `textInference`

Type of the task.

### [taskUUID](https://runware.ai/docs/models/minimax-m3#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID of the task.

### [text](https://runware.ai/docs/models/minimax-m3#response-text)

- **Type**: `string`
- **Required**: true

Generated text content.

### [reasoningContent](https://runware.ai/docs/models/minimax-m3#response-reasoningcontent)

- **Type**: `string`

The model's internal reasoning content, separated from the main response. Only present when `settings.splitThinking` is enabled.

### [cost](https://runware.ai/docs/models/minimax-m3#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

### [finishReason](https://runware.ai/docs/models/minimax-m3#response-finishreason)

- **Type**: `string`
- **Required**: true

The reason why the model stopped generating tokens.

**Possible values**: `stop` `length` `content_filter` `unknown`

### [usage](https://runware.ai/docs/models/minimax-m3#response-usage)

- **Path**: `usage.promptTokens`
- **Type**: `object (4 properties)`
- **Required**: true

Token usage statistics for the request.

#### [promptTokens](https://runware.ai/docs/models/minimax-m3#response-usage-prompttokens)

- **Path**: `usage.promptTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Number of tokens in the input prompt.

#### [completionTokens](https://runware.ai/docs/models/minimax-m3#response-usage-completiontokens)

- **Path**: `usage.completionTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Number of tokens generated in the response.

#### [totalTokens](https://runware.ai/docs/models/minimax-m3#response-usage-totaltokens)

- **Path**: `usage.totalTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Total number of tokens used (prompt + completion).

#### [thinkingTokens](https://runware.ai/docs/models/minimax-m3#response-usage-thinkingtokens)

- **Path**: `usage.thinkingTokens`
- **Type**: `integer`
- **Min**: `0`

Number of tokens used for internal reasoning. Billed separately.