
FLUX.1 Pro
BFL's flagship text-to-image model, API-only — the quality bar the open FLUX variants are distilled from.
About
FLUX.1 Pro launched on August 1, 2024 alongside Black Forest Labs itself, the startup founded by the core Stable Diffusion researchers. It is a 12-billion-parameter rectified-flow transformer noted at launch for photorealism, strong prompt adherence, and unusually good text rendering inside images — quality widely compared to Midjourney v6 and DALL-E 3. It powered xAI's Grok image generation at launch and was picked up by Mistral's Le Chat, and it remains the reference point the open FLUX.1 [dev] and [schnell] variants are distilled from. Pro is proprietary and API-only: no weights, served through BFL's own API and partners including fal.ai, Replicate, Together AI, and Freepik. Honest caveats: BFL has not disclosed the training data, which leaves open the same copyright questions hanging over most frontier image models, and per-image API pricing makes high-volume use a real budget line.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
BFL-reported: leading ELO scores on the Artificial Analysis image leaderboard (1.1 Pro); launch-era human-preference tests vs Midjourney v6 and DALL-E 3 with ~800 judges (provider-reported)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Flagship commercial model from BFL (founders: original Stable Diffusion creators); launched Aug 1, 2024 per VentureBeat.; Wide production adoption: powered xAI Grok image gen at launch, Mistral Le Chat (Nov 2024), API partners fal/Replicate/Together/Freepik.; BFL-reported leading ELO on Artificial Analysis image leaderboard (1.1 Pro).; Limitations: proprietary, no weights; training data undisclosed (copyright questions); API-only pricing..