# MiniMax Music 3 - Core.Today AI API

> MiniMax Music 3 via Fal.AI. Generates complete songs with vocals up to five minutes long from a music description (genre, mood, BPM, key, vocals, arrangement) and structured lyrics. Returns 44.1 kHz stereo WAV.

- **Provider**: MiniMax
- **Model ID**: minimax/music-3
- **Category**: Audio & TTS
- **Credits**: 280 per 60s track — billed per second of the requested duration (4.65 credits/s)
- **Speed**: Medium
- **Quality**: High

## Features
- Complete songs with vocals up to 300 seconds
- Structured lyrics with [verse], [chorus], [bridge] and more
- Structured captions for genre, BPM, key and arrangement control
- 44.1 kHz 16-bit stereo WAV output

## Use Cases
- Original songs for videos, ads and short-form content
- Demo tracks for songwriting and lyric drafts
- Jingles and branded theme songs
- Background songs for games and events

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Music description: style, mood, vocals, instrumentation and arrangement. For precise control, use a structured caption (genre, BPM, key, emotional progression, vocal details, section-by-section arrangement).
- **lyrics**: string - The lyrics to sing. Structure tags such as [intro], [verse], [pre-chorus], [chorus], [bridge], [instrumental], [solo] and [outro] must each be on their own line.

### Optional
- **duration**: number (default: 60) - Upper bound on the generated audio length in seconds (1-300). Billed per second of this requested value; the actual length is returned in the output.
- **num_inference_steps**: integer (default: 30) - Flow-matching steps per 8-second chunk (1-100). More steps improve quality at the cost of speed.
- **guidance_scale**: number (default: 1.7) - Classifier-free guidance scale of the flow-matching stage (0-20).
- **seed**: integer - Random seed for reproducibility. A random seed is used when omitted.

## Examples

### Acoustic pop song
A one-minute acoustic pop song driven by a structured caption and tagged lyrics.
```json
{
  "model": "minimax/music-3",
  "input": {
    "prompt": "Genre: acoustic pop. BPM: 96. Key: C major. Warm and intimate, building gently into the chorus. Vocals: soft female lead, close and breathy, light stacked harmonies in the chorus. Arrangement: fingerpicked guitar and soft piano; brushed drums and upright bass enter in the chorus.",
    "lyrics": "[verse]\nMorning light filtering through the pine\nEvery quiet street is yours and mine\n[chorus]\nSoftly the world begins to breathe",
    "duration": 60
  }
}
```

### Short upbeat jingle
A 20-second jingle — a shorter duration also lowers the charge.
```json
{
  "model": "minimax/music-3",
  "input": {
    "prompt": "Genre: upbeat synth-pop jingle. BPM: 120. Bright male and female duet, catchy hook.",
    "lyrics": "[chorus]\nOne key, every model\nBuild it today",
    "duration": 20
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Set duration close to the length you actually need — billing uses the requested value, not the shorter length the model may return.
- Put each structure tag ([verse], [chorus], ...) on its own line; text on the same line as a tag is dropped.
- Describe genre, BPM, key and vocal character in the prompt for more predictable results.
- For an instrumental passage, use an [instrumental] section in the lyrics.

## Documentation
https://fal.ai/models/minimax/music-3