MODEL ID moonshotai-kimi-k2-6

coming-soon

Kimi K2.6

by Moonshot AI April 20, 2026

Kimi K2.6 is Moonshot AI's latest flagship open model for coding, reasoning, multimodal understanding, and agentic execution. It is designed for long-horizon software tasks, reliable tool use, autonomous multi-step workflows, coordinated agent swarms, and visual understanding across image and video inputs in addition to text.

API Options

Platform-level options for task execution and delivery.

taskType string required value: textInference: Identifier for the type of task being performed

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

outputFormat 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 1 value

webhookURL string 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
PLATFORM

deliveryMethod string default: sync

Determines how the API delivers task results.

Allowed values 3 values

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

Learn more 1 resource

Task Polling
PLATFORM

includeCost boolean default: false: Include task cost in the response.

includeUsage boolean default: false: Include token usage statistics in the response.

numberResults integer min: 1 max: 4 default: 1: Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

Inputs

Input resources for the task (images, audio, etc). These must be nested inside the inputs object.

inputs » images images array of strings min items: 1: Array of image inputs (UUID, URL, Data URI, or Base64).

inputs » videos videos array of strings min items: 1: Array of video inputs (UUID, URL, or Base64).

Generation Parameters

Core parameters for controlling the generated content.

model string required value: moonshotai-kimi-k2-6: Identifier of the model to use for generation.

seed integer min: 0 max: 9223372036854776000: Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.

messages array of objects required min items: 1

Array of chat messages forming the conversation context.

Properties 2 properties

messages » role role string required: The role of the message author.

Allowed values 2 values

messages » content content string required 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.

settings » systemPrompt systemPrompt string min: 1 max: 50000: System-level instruction that guides the model's behavior and output style across the entire generation.

settings » temperature temperature 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.

settings » topP topP float min: 0 max: 1 step: 0.01 default: 1: Nucleus sampling parameter that controls diversity by limiting the probability mass. Lower values make outputs more focused, higher values increase diversity.

settings » frequencyPenalty frequencyPenalty float min: -2 max: 2 step: 0.01 default: 0: Penalizes tokens based on their frequency in the output so far. A value of 0.0 disables the penalty.

settings » maxTokens maxTokens integer min: 1 max: 1000000: Maximum number of tokens to generate in the response.

settings » minP minP float min: 0 max: 1 step: 0.01 default: 0: Minimum probability threshold. Tokens with probability below this value are excluded from sampling.

settings » presencePenalty presencePenalty float min: -2 max: 2 step: 0.01 default: 0: Encourages the model to introduce new topics. A value of 0.0 disables the penalty.

settings » promptCacheKey promptCacheKey string: Cache key for reusing prompt cache across requests. Requests sharing the same key reuse cached prompt processing.

settings » repetitionPenalty repetitionPenalty float min: 0.01 max: 5 step: 0.01 default: 1: Penalizes tokens that have already appeared in the output. A value of 1.0 disables the penalty.

settings » stopSequences stopSequences array of strings min: 1 max items: 16: Array of sequences that will cause the model to stop generating further tokens when encountered.

settings » thinkingLevel thinkingLevel string default: none: Controls the depth of internal reasoning the model performs before generating a response.

Allowed values 5 values

settings » topK topK integer min: 0 max: 999 default: 0: Top-K sampling parameter that limits the number of highest-probability tokens considered at each step.

toolChoice object

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

Examples 3 examples

Let the model decide (default):

"toolChoice": {
  "type": "auto"
}

Force a specific tool call:

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

Require any tool call:

"toolChoice": {
  "type": "any"
}

Properties 2 properties

toolChoice » type type string required

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

Allowed values 4 values

: The model decides whether to call a tool based on the conversation context. This is the recommended default.
: The model must call at least one tool but chooses which one. Useful when you always need structured output.
: The model must call the specific tool identified by name. Use this to force a particular function call.
: The model will not call any tool, even if tools are defined. Useful for forcing a text-only response.

toolChoice » name name string: Name of the specific tool the model must call. Required when type is tool.

tools array of objects min items: 1

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.

The search tool is executed server-side by the provider. You don't need to handle the tool result yourself.

The codeInterpreter tool is executed server-side by the provider. You don't need to handle the tool result yourself.

Examples 4 examples

Function tool, weather lookup:

"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:

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

Built-in code interpreter:

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

Multiple function tools:

"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"]
    }
  }
]

Properties 4 properties

tools » type type string required

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 3 values

: User-defined function tool. The model outputs the tool name and arguments. You execute the function locally and send results back.
: Built-in web search. The provider executes search server-side and enriches the response automatically.
: Built-in code execution sandbox (Python). The provider runs code server-side and returns results automatically.

tools » name name string max: 64: Unique function name. Required for function tools.

tools » description description string: Explanation of what the function does, used by the model to decide when to call it.

tools » schema schema object: JSON Schema object describing the function's input parameters.