Superseded - verified 2026-10-02
FLUX.1 (Aug 2024) was followed by FLUX.2 (Nov 2025) and FLUX 3 (2026-07-23), Black Forest Labs’ current family. BFL docs still call FLUX.2 “fully supported for production”.
FLUX family (Black Forest Labs)
by Black Forest Labs (BFL)
Text-to-image and editing models with open-weight and API tiers; now extended to video and audio.
Lineage
| Family | Date | Notes |
|---|---|---|
| FLUX.1 [schnell] / [dev] / [pro] | Aug 2024 | Schnell Apache 2.0; Dev non-commercial; Pro API-only. FLUX1.1 Pro (Oct 2024) |
| FLUX.1 Kontext | May 2025 | In-context generation and editing (Max, Pro, Dev) |
| FLUX.2 | Nov 2025 | Klein, Dev, Flex, Pro. Klein offered under Apache 2.0 or FLUX Non-Commercial; open weights on Hugging Face |
| FLUX 3 | 2026-07-23 | ”Multimodal flow model”; one model/API for text-to-video, image-to-video, audio and robotics action prediction per BFL docs; API and open weights per BFL blog |
| FLUX 3 Video / FLUX 3 Action | 2026-08-04 / 2026-09-23 | Variants listed on BFL blog |
BFL’s blog (bfl.ai/blog, 2026-10-02) confirms: FLUX 3 - Real World Models (2026-07-23), FLUX 3 Video Part 1 (2026-08-04), FLUX Video Upscale 2K/4K (2026-08-20) and FLUX 3 Action (2026-09-23, 7B open-weights world action model, first on the RoboLab benchmark per BFL).
Access status is reported inconsistently: at launch (2026-07-23/24) trade press said Video and Action started in gated early access, Image was “coming weeks”, the open-weight FLUX 3 Dev was promised later in the year, with no public licence. VentureBeat’s FLUX 3 Action article (2026-09) says FLUX 3 Video is generally available through BFL’s API (Draft HD to FHD tiers), while the Action weights’ final licence and any Action API availability are not yet announced. BFL’s docs landing page (fetched 2026-10-05) describes FLUX 3 as “one model, one API” covering image generation and editing, video with synchronised audio, audio and action prediction, and says FLUX.2 and FLUX.1 Kontext remain documented and supported (FLUX.2 [klein] under Apache 2.0 or non-commercial licence, FLUX.2 [dev] open weights). Open-weight FLUX 3 Dev availability and licence remain unverified; check bfl.ai/pricing.
Using FLUX
- Local: weights on Hugging Face,
diffusers, comfyui (ComfyUI lists Flux.1/2 support). - API: BFL directly, or hosts such as fal-dot-ai and kie-dot-ai.
- Strengths (commonly cited by users, opinion; not verified here): prompt adherence, typography, open-weight options.
Funding
TechCrunch (2025-12-01): USD 300 million Series B at a USD 3.25 billion valuation, co-led by Salesforce Ventures and AMP, with a16z, NVIDIA, General Catalyst and Temasek; total raised above USD 450 million. A search result also mentions a further USD 308 million round on 2026-06-03: single weak source, unverified.
Open items
- FLUX 3 Dev open weights and licence: not confirmed (docs 2026-10-05 do not state them).
- The June 2026 USD 308M round: unverified.
Related
midjourney, seedream-40, nano-banana, gpt-image-2-5
Sources (accessed 2026-10-02)
- https://bfl.ai/blog
- https://venturebeat.com/infrastructure/black-forest-labs-debuts-flux-3-action-an-open-weights-ai-robotics-model-that-tops-the-leaderboard-at-half-the-size-of-its-competition
- https://techcrunch.com/2025/12/01/black-forest-labs-raises-300m-at-3-25b-valuation
- https://novalogiq.com/2026/07/24/black-forest-labs-launches-flux-3-capable-of-generating-images-and-20-second-video-with-audio-but-in-limited-release-to-start/
- https://docs.bfl.ai/ (accessed 2026-09-30, re-fetched 2026-10-05)
- https://en.wikipedia.org/wiki/Flux_(text-to-image_model) (accessed 2026-09-30)