FLUX.2 Dev
Open-weight 32-billion parameter AI image model from Black Forest Labs. Self-host or fine-tune with full access to weights, multi-reference conditioning, and unified text-to-image plus editing.
Explore creations made with FLUX.2 Dev
FLUX.2 Dev from Black Forest Labs puts professional-grade image generation in the hands of developers and researchers. This 32-billion-parameter open-weight model integrates text-to-image generation and image editing into a single unified architecture. With multi-reference conditioning built in and no separate modules required, FLUX.2 Dev offers the flexibility that serious developers need while delivering quality that rivals closed-source alternatives.
Open Weights
Full access to model weights for customization, fine-tuning, and integration into your own pipelines and products.
Multi-Reference Conditioning
Support for multiple reference images without requiring separate modules or complex pipelines.
Unified Architecture
Text-to-image generation and image editing in one model, simplifying deployment and reducing infrastructure complexity.
Local Deployment
Run locally using BFL's reference code or optimized FP8 implementations developed with NVIDIA and ComfyUI. 40% lower VRAM requirements with FP8 quantization.
- Building custom AI image applications
- Research and experimentation with diffusion models
- Creating specialized fine-tuned variants
- Self-hosted image generation for privacy-sensitive applications
FLUX.2 Dev is the choice for developers who need control. The open weights mean you can fine-tune for specific use cases, deploy on your own infrastructure, and build products without API dependencies. Benchmarks show it leading open-weight alternatives with a 66.6% win rate in text-to-image tasks. If you're building something that requires more than API calls to a black box, FLUX.2 Dev gives you the foundation to build it right.




