ICYMI: Black Forest Labs opens FLUX 3 Video early access
FLUX 3 Video is in early access with 20-second video generation and native audio from text, images, keyframes, or clips.
TestingCatalog says Black Forest Labs is presenting it as the first available piece of a jointly trained image, video, and audio foundation model. The model can extend clips, preserve central elements across scenes, generate multilingual dialogue, and chain shots into longer sequences. The company cites early preference-test wins over several rival video systems, while noting that both the model and evaluation setup are still in development. FLUX 3 Image access is expected later, with APIs, private weights, and an open-weight Dev backbone planned. TestingCatalog's note
TestingCatalog says Black Forest Labs is presenting it as the first available piece of a jointly trained image, video, and audio foundation model. The model can extend clips, preserve central elements across scenes, generate multilingual dialogue, and chain shots into longer sequences. The company cites early preference-test wins over several rival video systems, while noting that both the model and evaluation setup are still in development. FLUX 3 Image access is expected later, with APIs, private weights, and an open-weight Dev backbone planned. TestingCatalog's note
score 6