Flux.2 Pro - High-Quality Image Generation with Multi-Reference Control

Flux 2 Pro is Black Forest Labs' advanced image generation and editing model powered by a latent flow matching architecture combining a 24-billion parameter vision-language model with a rectified flow transformer. It creates high-quality images from text prompts and edits existing images with natural language instructions.

Text rendering with legible typography, photorealistic details with sharp textures and natural lighting, character consistency across multiple images, and structured JSON prompting for precise scene control.

Organize by importance: subject first, then action, style, and context. Example: 'A sleek silver sports car racing along a coastal highway at sunset, high-dynamic-range, hyper-realistic.' Important: Never use negative prompts like 'no text' or 'no extra fingers'—they may cause the opposite effect. Instead, describe what you want to see.

Yes. Upload an image and describe changes in natural language. Reference specific parts or other reference images by index. Example: 'Replace the background with the beach in image 3' or 'Make the woman wear a red jacket.'

Structure your prompt as JSON with fields like scene, subjects, style, lighting, camera (angle, distance, focal length, aperture, ISO), and color_palette (hex codes or color names) for precise control over every aspect.

Upload multiple images to guide style, composition, or specific elements. Total input is limited to 9 megapixels on the API. The model uses these to maintain consistency across generations.

Use hex codes in prompts for exact color matching—ideal for brand work and marketing materials requiring specific color accuracy. The model edits images at resolutions up to 4 megapixels while maintaining detail and coherence.

Yes. Free credits are included for new users. Buy additional credits or subscribe for more usage. Unused credits don't expire, and failed jobs are auto-refunded. See Pricing for model plans.