> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.empiriolabs.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.empiriolabs.ai/_mcp/server.

# Stable Audio 3 Medium

> Stability AI text-to-audio model for music and sound effects, 44.1 kHz stereo clips up to 380 seconds, with length, step and seed controls.

![Stable Audio 3 Medium](https://media.empiriolabs.ai/model-logos/stable-audio-3-medium.png)

[Stability AI](/providers/stability) · Audio Generation

`POST /v1/audio/generations`

Stability AI text-to-audio model for music and sound effects, 44.1 kHz stereo clips up to 380 seconds, with length, step and seed controls.

## At a glance

| Field               | Value                                                        |
| ------------------- | ------------------------------------------------------------ |
| Model id            | `stable-audio-3-medium`                                      |
| Model release date  | 2026-05-20                                                   |
| Input modalities    | Text                                                         |
| Output modalities   | Audio                                                        |
| Context window      | -                                                            |
| Weight precision    | FP16                                                         |
| Features            | music_generation, text_to_audio, sound_effects, seed_control |
| Native inference    | Yes                                                          |
| New                 | Yes                                                          |
| Supported endpoints | `POST /v1/audio/generations`                                 |
| Alternate model ids | `stabilityai/stable-audio-3-medium`                          |

## Pricing

| Charge     | Spec        | Rate    |
| ---------- | ----------- | ------- |
| Generation | per request | \$0.035 |

## Example request

```bash
curl https://api.empiriolabs.ai/v1/audio/generations \
  -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"model": "stable-audio-3-medium", "prompt": "rain on a tin roof", "audio_duration": 8}'
```

## Parameters

| Parameter             | Type    | Required | Default | Description                                                                                       |
| --------------------- | ------- | -------- | ------- | ------------------------------------------------------------------------------------------------- |
| `prompt`              | string  | yes      | -       | The music or sound to generate.                                                                   |
| `audio_duration`      | number  | no       | `30`    | Clip length in seconds, from 1 to 380. · Range: 1 – 380                                           |
| `num_inference_steps` | number  | no       | `8`     | Sampling steps, from 4 to 8. · Range: 4 – 8                                                       |
| `seed`                | number  | no       | -       | Random seed. Leave unset for a random one. · Range: 0 – 2147483647                                |
| `format`              | enum    | no       | `"mp3"` | Audio file format of the generated clip. · Allowed: `mp3`, `wav`, `flac`, `ogg`                   |
| `response_format`     | enum    | no       | `"url"` | Return a signed media URL by default, or base64 JSON when requested. · Allowed: `url`, `b64_json` |
| `return_base64`       | boolean | no       | false   | Compatibility alias for response_format=b64_json.                                                 |

## Notes

Generates music and sound effects from text: 44.1 kHz stereo, from 1 to 380 seconds per clip. Powered by Stability AI.

**Input**

* A text prompt describing the music or sound: genre, instruments, mood, tempo, or the effect.
* `audio_duration` sets the length in seconds, from 1 to 380. The default is 30.
* `num_inference_steps` from 4 to 8. The default is 8.
* `seed` repeats a result. Negative prompts, lyrics and audio inputs are not supported.

**Output**

* MP3 by default, or WAV, FLAC or Ogg with `format`.
* A signed URL by default, or base64 with `response_format` set to `b64_json`.

**Billing**

* Billed per clip, whatever its length. A request that fails is not billed.

**Weights**

* FP16.

---

*Machine-readable schema:* `GET https://api.empiriolabs.ai/v1/models/stable-audio-3-medium`.