# FLUX 2 Max - Core.Today AI API

> The highest fidelity image model from Black Forest Labs. Best-in-class prompt following and the most consistent editing in the FLUX.2 lineup — preserves colors, lighting, faces, text, and objects across edits with up to 8 reference images.

- **Provider**: Black Forest Labs
- **Model ID**: black-forest-labs/flux-2-max
- **Category**: Image Generation
- **Credits**: 160 per image at 1 MP — varies by resolution (0.5MP 130, 2MP 230, 4MP/match_input 370, custom 390)
- **Speed**: Fast
- **Quality**: Ultra

## Features
- Highest editing consistency in the FLUX.2 lineup — preserves identity, colors, lighting, and text
- Best-in-class prompt following for both short and long prompts
- Multi-reference editing with up to 8 input images
- Systematic product color variations controlled by hex codes in the prompt
- New 3D views of a scene from a single image or small reference set
- Nearly as fast as FLUX.2 [pro] — up to 3x faster than competing models of similar quality

## Use Cases
- Top-tier product imagery from zero to publish-ready in minutes
- Transforming low-quality phone photos into publishing-grade product shots
- Cinematic key frames for animation with preserved characters and emotions
- Dozens of ad variants with the same face and consistent product mockups
- Interior design, 3D reconstruction, and food imagery final polish

## 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 - Text prompt for image generation

### Optional
- **input_images**: array (default: []) - Up to 8 reference images (jpeg, png, gif, or webp) for image-to-image generation
- **aspect_ratio**: string (default: 1:1) - Aspect ratio. 'match_input_image' follows the first input image; 'custom' uses width/height
  Options: match_input_image, custom, 1:1, 16:9, 3:2, 2:3, 4:5, 5:4, 9:16, 3:4, 4:3
- **resolution**: string (default: 1 MP) - Output resolution in megapixels. 2 MP or below recommended; max image size is 2048x2048. Billing scales with resolution
  Options: match_input_image, 0.5 MP, 1 MP, 2 MP, 4 MP
- **width**: integer - Width (256-2048, multiple of 16). Only used when aspect_ratio=custom
- **height**: integer - Height (256-2048, multiple of 16). Only used when aspect_ratio=custom
- **safety_tolerance**: integer (default: 2) - Safety tolerance: 1 is most strict, 5 is most permissive
- **seed**: integer - Random seed for reproducible generation
- **output_format**: string (default: webp) - Output image format
  Options: webp, jpg, png
- **output_quality**: integer (default: 80) - Save quality 0-100 (not relevant for png)

## Examples

### Photorealistic Scene with Text
Complex scene with accurate in-image text rendering
```json
{
  "model": "black-forest-labs/flux-2-max",
  "input": {
    "prompt": "A photorealistic wide shot of a minimalist digital billboard in a forest of purple and lavender-hued trees. Centered text reads \"FLUX.2 [max]\" in a modern sans-serif font. Soft daylight with a purple color cast, cinematic composition",
    "resolution": "1 MP",
    "aspect_ratio": "1:1"
  }
}
```

### Multi-Reference Product Edit
Consistent product mockup across a new environment
```json
{
  "model": "black-forest-labs/flux-2-max",
  "input": {
    "prompt": "Place the product from the reference image on a marble kitchen counter with soft morning light, keeping the label text and colors exactly as shown",
    "input_images": [
      "https://example.com/product.jpg"
    ],
    "resolution": "2 MP",
    "aspect_ratio": "4:5"
  }
}
```

## 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
- Billing scales with output resolution — iterate at 1 MP, then render finals at 2-4 MP
- 2 MP or below is recommended; at 4 MP non-square ratios may not reach the full pixel count (2048x2048 cap)
- Use hex color codes in the prompt for systematic product color variations
- When editing, state what must stay unchanged — the model preserves faces, text, and lighting exceptionally well
- Use 'match_input_image' for both aspect_ratio and resolution to keep an edit dimensionally identical to the source

## Documentation
https://replicate.com/black-forest-labs/flux-2-max