GET

/v1/schemas/minimax/chat/anthropic/v1/messages

Accept:

kind inputactivity chathash 67d78b89…extracted Nov 8, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `c41e967945ea1159466901032e21618b0da87cb1063ee2f3c244cc6e998b8752` { "fetchedAt": 1794115838, "extractorVersion": "2", "verifiedAt": null }

Request body

13 properties · 2 required · 6 $defs

property	type	description	constraints
model*	enum	Model ID. `MiniMax-M3.1-Flash` and `MiniMax-M3` are multimodal models with native support for text, image, and video input, alongside tool use and thinking content blocks. `MiniMax-M3.1-Flash` additionally always thinks and supports tunable thinking depth via …	MiniMax-M3.1-Flash \| MiniMax-M3
messages*	Message[]	Conversation history. `MiniMax-M3.1-Flash` and `MiniMax-M3` support text, image, video, tool use, tool result, and thinking content blocks.
service_tier	enum	Service tier for request admission. Supported values are `standard` and `priority`. If omitted, the request uses the `standard` tier. The `priority` [price](https://platform.minimax.io/docs/pricing/overview#models/text) is 1.5 times the `standard` price and en…	standard \| priority · default: "standard"
system	string \| object[]	Set the role and behavior of the model.
stream	boolean	Whether to use streaming output, defaults to `false`. When set to `true`, the response will be returned in chunks	default: false
max_tokens	integer	Specifies the upper limit for generated content length (in tokens). For MiniMax-M3.1-Flash and MiniMax-M3 the recommended value is 131072 (128K) and the maximum is 524288 (512K). Content exceeding the limit will be truncated. If generation stops due to `length…	n ≥ 1 · format: int64
temperature	number	Not supported yet; any value passed has no effect. `MiniMax-M3.1-Flash` and `MiniMax-M3` always use 1.	format: double · default: 1
top_p	number	Not supported yet; any value passed has no effect. `MiniMax-M3.1-Flash` and `MiniMax-M3` always use 0.95. If passed, the value must be within (0, 1].	0 ≤ n ≤ 1 · format: double · default: 0.95
tools	Tool[]	Tool definitions for Anthropic-compatible tool use.
tool_choice	ToolChoice	Tool selection strategy. Only `auto` is supported; the model decides whether to call tools.
thinking	object	Controls thinking behavior. The default depends on the model, so no schema-level default is declared. - `MiniMax-M3.1-Flash`: always thinks. If `type` is sent it must be `adaptive`; `disabled` returns HTTP 400. Use `output_config.effort` to tune thinking depth…
output_config	object	Output configuration. Use `effort` to tune thinking depth for `MiniMax-M3.1-Flash`.
metadata	object	Request metadata. user_id is recommended for end-user-level aggregation, rate limiting, and billing analysis.

Types

6 $defs

CacheControlobject

Prompt cache marker. Not supported by MiniMax-M3 or MiniMax-M3.1-Flash.

type*

enum

Prompt cache marker.

ephemeral

MediaSourceobject

Media source for image or video blocks.

type*	enum	Media source type.	base64 \| url
media_type	string	Required for base64 input, for example image/png or video/mp4.
data	string	Base64-encoded media bytes.
url	string	Public URL. Videos can also use a Files API reference in the form `mm_file://{file_id}`. To obtain a `file_id`, first upload the video via the [Upload File](/api-reference/file-management-upload) API.
detail	enum	Understanding detail level. Defaults to default.	low \| default \| high · default: "default"
fps	number	Video frame sampling rate. Defaults to 1. Range [0.2, 5]. - Higher values: more sensitive to scene changes, but higher token cost and slower response. - Lower values: lower token cost and faster response, but less sensitive to scene changes.	0.2 ≤ n ≤ 5 · default: 1
max_long_side_pixel	integer	Longest-side pixel limit for the image or video frames.	n ≥ 1

Messageobject

role*	enum	Role of the message sender. `MiniMax-M3.1-Flash` and `MiniMax-M3` use alternating user / assistant messages.	user \| assistant \| user_system \| group \| sample_message_user \| sample_message_ai
content*	string \| RequestContentBlock[]	Message content. `MiniMax-M3.1-Flash` and `MiniMax-M3` support text, image, video, tool use, tool result, and thinking content blocks.

Content block accepted in request messages. text and tool blocks are supported by all models; image and video blocks are supported by `MiniMax-M3.1-Flash` and `MiniMax-M3` only. thinking is only used when echoing a prior assistant turn back to the model.

type*	enum	Content block type: - text: Text content - image: Image input, `MiniMax-M3.1-Flash` and `MiniMax-M3` only - video: Video input, `MiniMax-M3.1-Flash` and `MiniMax-M3` only - tool_use: Echoed assistant tool call from a previous turn - tool_result: Tool execution…	text \| image \| video \| tool_use \| tool_result \| thinking · +1 more
text	string	Text content when type=text.
source	MediaSource	Image or video source when type=image or type=video. `MiniMax-M3.1-Flash` and `MiniMax-M3` only.
id	string	Tool call ID when echoing type=tool_use.
name	string	Tool name when echoing type=tool_use.
input	object	Tool input when echoing type=tool_use.
tool_use_id	string	ID of the matching tool_use block when type=tool_result.
content	string \| RequestContentBlock[]	Tool execution result, either a string or an array of text/image content blocks.
thinking	string	Echoed assistant thinking content when type=thinking.
signature	string	Signature for echoed thinking content. Return unchanged in later turns.
cache_control	CacheControl	Prompt cache marker. Not supported by MiniMax-M3 or MiniMax-M3.1-Flash.

Toolobject

Tool definitions for Anthropic-compatible tool use.

name*	string	Tool name.
description	string	Tool description.
input_schema*	object	Tool input parameters JSON Schema.
cache_control	CacheControl	Prompt cache marker. Not supported by MiniMax-M3 or MiniMax-M3.1-Flash.

ToolChoiceobject

Tool selection strategy. Only `auto` is supported; the model decides whether to call tools.

type*

enum

Tool choice type.

auto

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"minimax","endpointId":"anthropic/v1/messages","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}