Overview

Generate speech from text with MiniMax speech models, using English or Mandarin system voices. Supports MP3, WAV, FLAC, and PCM output with adjustable speaking speed.

Currently available model: speech-2.8-turbo.

Pricing

Model (`task_type`)	Price (USD)
`speech-2.8-turbo`	$0.09 per 1,000 billed characters

Charges use the actual billed character count, which can differ from the input text length. Check final usage and billing records; a failed task does not necessarily mean no charge was incurred.

Notes

text: 1–9,999 Unicode characters; valid UTF-8 and not entirely whitespace.

voice: use an exact system voice ID listed in Body Params, including case, spaces, and parentheses. Custom voices and voice cloning are not supported.

speed: 0.5–2, default 1.

response_format: mp3 (default), wav, flac, or pcm. PCM is raw audio and requires a compatible decoder.

Streaming is not supported. Omit stream and stream_format, or set them to false and audio.

Only service_mode: public is supported. Unlisted input fields are rejected.

Request

Header Params

Body Params application/jsonRequired

Examples

English MP3Mandarin WAV

{
  "model": "minimax",
  "task_type": "speech-2.8-turbo",
  "input": {
    "text": "Hello, welcome to this audio demonstration.",
    "voice": "English_expressive_narrator",
    "response_format": "mp3",
    "speed": 1
  }
}

Request Code Samples

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

Responses

🟢200OK

application/json

Task accepted. Save data.task_id and query its status; generation is asynchronous.

Bodyapplication/json

Example

Illustrative creation response. Query the returned task ID for completion.

{
  "code": 200,
  "message": "success",
  "data": {
    "task_id": "00000000-0000-4000-8000-000000000001",
    "model": "minimax",
    "task_type": "speech-2.8-turbo",
    "status": "pending",
    "config": {
      "service_mode": "public",
      "webhook_config": {
        "endpoint": "",
        "secret": ""
      }
    },
    "input": {
      "text": "Hello, welcome to this audio demonstration.",
      "voice": "English_expressive_narrator",
      "response_format": "mp3",
      "speed": 1
    },
    "output": null,
    "meta": {
      "created_at": "2026-10-16T08:00:00Z",
      "started_at": "0001-01-01T00:00:00Z",
      "ended_at": "0001-01-01T00:00:00Z",
      "usage": {
        "type": "point",
        "frozen": 0,
        "consume": 0
      }
    },
    "error": {
      "code": 0,
      "message": "",
      "detail": null
    }
  }
}

MiniMax Speech

Overview#