> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kymaapi.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MiniMax Speech Turbo

> MiniMax low-latency voice on Kyma. ~2.2× cheaper than ElevenLabs Flash v2.5. Best for real-time voice agents.

## Overview

`minimax-speech-turbo` is MiniMax's low-latency text-to-speech tier. Multilingual, fast time-to-first-byte, cheapest voice on Kyma. Right for bulk TTS, real-time voice agents, conversational AI, and high-throughput pipelines.

## Specs

| Field        | Value                                               |
| ------------ | --------------------------------------------------- |
| Model ID     | `minimax-speech-turbo`                              |
| Creator      | MiniMax                                             |
| Best for     | Real-time agents, conversational AI, bulk narration |
| Max input    | 5000 characters per request                         |
| Pricing mode | Per character                                       |

## Pricing

|                                |    Cost   |
| ------------------------------ | :-------: |
| Per 1K chars                   |  \$0.090  |
| Typical sentence (\~150 chars) | \~\$0.014 |

## Use this when

* You're building a real-time voice agent and TTFB matters.
* You're processing high volume (chatbot replies, IVR menus, bulk narration).
* You need a budget-friendly default that still sounds production-acceptable.

## Pick something else when

* You need richer expressivity for storytelling → [`minimax-speech-hd`](/models/minimax-speech-hd) (\$0.140/1K char).
* You're cutting hero brand spots → [`eleven-multilingual-v2`](/api-reference/audio-speech).

## Example

```bash theme={null}
curl -X POST https://kymaapi.com/v1/audio/speech \
  -H "Authorization: Bearer $KYMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-speech-turbo",
    "input": "Hi, how can I help you today?",
    "voice_id": "female-shaonv"
  }' \
  --output reply.mp3
```

## See also

* [`POST /v1/audio/speech`](/api-reference/audio-speech) — endpoint reference
* [MiniMax Speech HD](/models/minimax-speech-hd) — production-quality tier