FLUX (Black Forest Labs)

FLUX is a family of text-to-image generation models created by Black Forest Labs, a company founded in 2024 by Robin Rombach, Andreas Blattmann, and Patrick Esser — the core researchers behind Stable Diffusion. The company set out to push the boundaries of open-source image generation, and within months of launch, FLUX models became among the most widely deployed image generation systems in the world.

The original FLUX.1 lineup introduced three tiers: FLUX.1 Schnell (Apache 2.0 licensed and freely available for commercial use), FLUX.1 Dev (non-commercial open weights on Hugging Face), and FLUX.1 Pro (API-only, proprietary). Each tier offered a different tradeoff between speed, quality, and accessibility — Schnell being the fastest and most permissive, Pro delivering the highest fidelity.

The FLUX.2 generation, launched in late 2025, expanded the lineup significantly. FLUX.2 [pro] generates images at resolutions up to 4 megapixels with sharp text rendering, accurate color matching, and consistent character identity across multiple outputs — features that have historically been weak points in AI image generation. FLUX.2 [max] pushes quality further still, while [flex] and [klein] cater to different cost and speed profiles. The newest addition, FLUX Kontext, enables conversational image editing: instead of starting from scratch with each prompt, you build on a generated image through follow-up instructions, adjusting specific elements while preserving the rest.

FLUX is available through Black Forest Labs' own API with pay-as-you-go pricing based on megapixels: FLUX.2 [pro] starts at $0.03/MP, [max] at $0.07/MP, and [klein] at roughly $0.014–$0.015 per image. The open-weight models can be run locally or through platforms like Hugging Face, Replicate, and Fal.ai, often at lower cost than the official API.

The FLUX models have earned a reputation for exceptional prompt adherence — accurately rendering complex scenes, specific styles, and detailed text within images — along with photorealism that rivals or exceeds commercial competitors. They are used extensively in commercial creative workflows, marketing asset generation, game concept art, and AI-powered design tools.

**Who it's for:** Individual creators use the open-weight models locally or via third-party platforms for maximum flexibility and low cost. Developers and studios building image generation into products use the API for scalable, high-quality output. Businesses generating photorealistic assets at scale — ad creatives, product imagery, visual branding — find the quality-per-dollar compelling.

**Getting started:** The easiest entry point is through a third-party interface like Fal.ai, Replicate, or platforms that embed FLUX models. Direct API access requires creating an account at bfl.ai and purchasing credits. Local deployment of the open-weight Schnell model requires a capable GPU (16GB VRAM recommended for full-quality output).