[AINews] Black Forest Labs FLUX 3 - Multimodal Flow Models that beat Seedance 2.0, Gemini Omni and Grok Imagine, and FLUX-mimic video-action robotics model
Black Forest Labs says FLUX 3 is one unified model for image, video, audio, and action prediction.
Latent Space frames the launch as BFL’s long-promised move from FLUX image generation into video, with native audio generation and modes including text-to-video, image-to-video, video-to-video, keyframes, dialogue, typography, and multi-shot chaining. The piece says BFL is claiming strong performance against other frontier media systems, with an open-weights Dev version planned. It also highlights FLUX3-mimic, a robotics partner model built on the FLUX 3 backbone for video-action control and factory testing.
Latent Space's note
Latent Space frames the launch as BFL’s long-promised move from FLUX image generation into video, with native audio generation and modes including text-to-video, image-to-video, video-to-video, keyframes, dialogue, typography, and multi-shot chaining. The piece says BFL is claiming strong performance against other frontier media systems, with an open-weights Dev version planned. It also highlights FLUX3-mimic, a robotics partner model built on the FLUX 3 backbone for video-action control and factory testing.
Latent Space's note
score 8