Flux 3 Max

• 🎬 Image-to-video: animates a still into a clip from a prompt
• 🔊 Native synchronized audio — speech, effects and ambience generated with frames
• 📏 Up to 20 seconds, Full HD 1920×1088 at 24 fps
• 🧩 One multimodal model generating video with synchronized audio
• 🔧 Keyframe pinning, clip continuation and prompt-based video editing
• ⚡ Draft mode returns fast previews before full-quality rendering
• 🏢 Black Forest Labs, Freiburg, Germany; available via BFL API
• 🌐 Multilingual dialogue with lipsync alongside the visuals

Black Forest Labs

Black Forest Labs is a generative AI company based in Freiburg im Breisgau, Germany, founded by former members of Stability AI. The lab is best known for developing the Flux family of text-to-image models, which generate images from natural language prompts…

Flux 3 Max (image-to-video) is the higher-quality tier of Black Forest Labs' Flux 3 video stack, taking a still image plus a text prompt and producing a moving clip with sound. It shares the Flux 3 Video backbone with Flux 3 Max (text-to-video) and the earlier Flux 3 image-to-video release, and sits in the same generation as the still-image Flux 3 and its editing counterpart Flux 3.

The jump from the Flux 2 line is structural. Flux 2 Max and Flux 2 Pro were image-only systems. Flux 3 is described by BFL as one multimodal model that generates video with synchronized audio — from a text prompt, from pinned keyframes, or by continuing an existing clip — up to 20 seconds at Full HD (1920×1088 for 16:9) and 24 fps in a single request. Audio is produced alongside the frames rather than added later, covering multilingual speech with lipsync plus effects and ambience. Nothing in the Flux 2 family had a video or audio counterpart.

Practical controls include start, end and ordered intermediate keyframes, multi-shot scenes and camera angles in one video, and text-prompted editing of an existing clip where unmentioned elements stay as shot. A draft mode returns a quick low-cost preview that can be re-submitted for a full-quality render of the same shot. The model is offered through the BFL API alongside the rest of the Flux 3 generation.