# FLUX.2 Klein 4B - Core.Today AI API

> Very fast image generation and editing model. 4-step distilled, sub-second inference for production and near real-time applications.

- **Provider**: Black Forest Labs
- **Model ID**: black-forest-labs/flux-2-klein-4b
- **Category**: Image Generation
- **Credits**: 2 per image at 1MP (default); 2 credits at 0.25-0.5MP, 4 at 2MP, 8 at 4MP
- **Speed**: Fast
- **Quality**: Standard

## Features
- 4-step distilled model — sub-second inference
- Supports both text-to-image and image-to-image editing (up to 5 reference images)
- Selectable output resolution from 0.25MP to 4MP
- Optional go_fast mode for additional speed
- Billed by output megapixels, so smaller previews cost less

## Use Cases
- Near real-time preview generation in interactive apps
- High-throughput batch image editing pipelines
- Low-cost draft generation before a final high-resolution render
- Fast iteration on prompts and reference-image edits

## 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
- **images**: array (default: []) - Input images for image-to-image generation (max 5)
- **aspect_ratio**: string (default: 1:1) - Aspect ratio for the generated image. Use match_input_image to match the first input image
  Options: 1:1, 16:9, 9:16, 3:2, 2:3, 4:3, 3:4, 5:4, 4:5, 21:9, 9:21, match_input_image
- **output_megapixels**: string (default: 1) - Resolution of the output image in megapixels. Higher values cost more credits
  Options: 0.25, 0.5, 1, 2, 4
- **seed**: integer - Random seed. Set for reproducible generation
- **go_fast**: boolean (default: false) - Run faster predictions with additional optimizations
- **output_format**: string (default: jpg) - Format of the output images
  Options: webp, jpg, png
- **output_quality**: integer (default: 95) - Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs.
- **disable_safety_checker**: boolean (default: false) - Disable safety checker for generated images.

## Examples

### Fast 1MP Generation (Default)
Default 1-megapixel text-to-image generation
```json
{
  "model": "black-forest-labs/flux-2-klein-4b",
  "input": {
    "prompt": "A surreal underwater or cosmic garden scene with bioluminescent flora against a deep navy background",
    "output_megapixels": "1",
    "aspect_ratio": "1:1"
  }
}
```

### Low-Cost Preview
0.25-megapixel draft for quick prompt iteration
```json
{
  "model": "black-forest-labs/flux-2-klein-4b",
  "input": {
    "prompt": "A neon-lit cyberpunk alley at night, concept sketch",
    "output_megapixels": "0.25"
  }
}
```

## 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
- output_megapixels directly sets the credit tier — use 0.25-0.5MP for fast, cheap previews and 2-4MP only for final renders
- Sub-second inference makes this a good fit for interactive or near real-time generation UIs
- Supply up to 5 reference images to guide editing via the images parameter
- go_fast trims latency further at a very small quality trade-off

## Documentation
https://replicate.com/black-forest-labs/flux-2-klein-4b