Black Forest Labs

@bfl_ai

Powering visual intelligence everywhere. Creators of FLUX models. join us: https://bfl.ai/careers

following
followers

fal launches FLUX 3 Max video model built with Black Forest Labs

fal reports a 12.45-second generation time for a 10-second clip in one comparison, not a guaranteed wait.

12 Sources, 1d ago, updated 1d ago

Science

FLUX 3 Max claims a No. 1 image-to-video tie and 22× Seedance 2.5's speed

Black Forest Labs says fal adapted FLUX 3 into FLUX 3 Max, which can generate 10-second videos in 12 seconds.

5 Sources, 1d ago, first seen 1d ago

FLUX 3 Max unveiled with a three-week 75%-off offer on October 19, 2026

Fal says the model, developed with @bfl_ai, generates a 10-second 720/768p video in about 12 seconds.

24 Sources, 4h ago, updated 4h ago

FLUX 3 Image Klein workflows updated to fix some issues

BFL AI recommends ComfyUI v0.41.3 and says it has also updated its Hugging Face Space.

1 Source, 3d ago, first seen 3d ago

FLUX 3 Image Klein launches with open weights and a 5.6B-parameter model

BFL says it runs on a single consumer GPU and offers native 4K generation, bounding boxes and pixel-perfect editing.

31 Sources, 1d ago, updated 1d ago

Technology

FLUX 3 Image Klein arrives with open weights

BFL says the 5.6-billion-parameter model runs on a single consumer GPU and offers native 4K.

4 Sources, 5d ago, updated 5d ago

FLUX 3 Video is now up to 2x faster and costs 50% less for native 2K and 4K generation

@bfl_ai says optimizations reduced the compute needed to serve FLUX 3 Video, with pricing and speed improvements reflecting that change.

5 Sources, 5d ago, updated 5d ago

FLUX 3 demos and a FLUX-mimic robot planned for GTC Berlin

BFL AI says FLUX 3 Action will be at the Nebius booth, with Image and Video at NVIDIA’s.

1 Source, 7d ago, first seen 7d ago

FLUX 3's claimed No. 1 ranking on the Physics-IQ benchmark

BFL AI says the benchmark tests whether video models can predict what follows the start of real filmed experiments.

6 Sources, 14d ago, updated 14d ago

Image creation promises up to 10 references and edits that leave other pixels unchanged

The bfl_ai account also highlights text-to-image and image-to-image creation, with generation at up to 4K.

19 Sources, 17d ago, updated 17d ago

FLUX 3 Action launches as an open-weight model for robot control

BFL says teams can fine-tune the 7-billion-parameter model for specific robots and tasks, with native Hugging Face LeRobot integration and deployment on Nvidia Jetson.

3 Sources, 27d ago, updated 27d ago