FLUX.1

FLUX.1 is a Text-to-Image model family developed by Black forest Labs, a team with deep roots in the open image-generation community. Instead of shipping a single model, the project offers three variants that target different audiences and workloads, so users can pick the balance of quality, speed, and licensing that suits them.

What it does

FLUX.1 turns written Prompts into images. The flagship variant, FLUX.1 [pro], is positioned for the highest visual fidelity, strong prompt adherence, and varied ouTPUts, and it is offered through an API with partner access. FLUX.1 [dev] is an open-weight, guidance-distilled model intended for non-commercial use, with quality and prompt following close to the pro tier but at higher efficiency. FLUX.1 [schnell] is the fastest option, released under an Apache 2.0 license and AImed at local development and Personal experimentation.

Key capabilities

• High-quality text-to-image synthesis with strong prompt following and fine detail in the pro tier.
• Open weights for the dev and schnell variants, enabling local runs, fine-tuning experiments, and research.
• Multiple access routes: hosted dEMOs and APIs via partners, model weights on Hugging Face, and inference code on GitHub.
• Speed-oriented generation in schnell for rapid iteration and personal projects.
• Permissive licensing on schnell, while dev cARRies non-commercial restrictions that users should verify.

Who it is for

FLUX.1 suits several groups. designers, illustrators, and marketers who need polished visuals can use the pro tier through an API. Developers and researchers can download open weights to build custom pipelines, test ideas, or integrate Image Generation into APPs. Hobbyists and students can start with schnell locally, since it is the lightest and most openly licensed variant. Teams evaluating models for commercial products should review each variant's license, because terms differ and may change over time.

How to get started

1. Decide which variant matches your needs: pro for top quality, dev for open-weight non-commercial work, or schnell for fast local use.
2. For pro, obtain API access through the official channels or participating partners; for dev and schnell, download weights from the model hub and set up the provided inference code.
3. Prepare a clear prompt describing subject, style, lighting, and composition, then generate and refine results iteratively.
4. Check hardware requirements before local runs, since larger variants benefit from capable GPUs.

Pricing and access caveat

Pricing, availability, and licensing details may vary by provider and can change without notice. The pro tier is typically billed through API usage or partner platforms, while open-weight variants are free to download subject to their licenses. Users should confirm current terms on the official site before relying on FLUX.1 for production or commercial work.