Describe what you want to generate
FLUX 3 Video – Native-Audio Multimodal Video Generation
Discover the confirmed capabilities of Black Forest Labs' first FLUX video model: native audio, clips up to 20 seconds, text-to-video, image-to-video, video references, keyframe control, multilingual dialogue, and multi-shot chaining. FLUX 3 Video is coming soon to Flux2Pro.
FLUX 3 Video Preview
Current preview clips illustrating native-audio video, character scenes, motion, and multi-shot potential. Results are preliminary and may differ from the final release.
Create Expressive Action Scenes
FLUX 3 Video is designed to interpret detailed direction for subjects, movement, camera behavior, environment, and synchronized sound in one generation.
Prompt-Directed Motion
Describe actions, timing, camera movement, atmosphere, and physical events in a structured scene brief.
Native Event Audio
Generate sound together with the visual action so physical events can feel connected to what is on screen.
Longer Single Clips
BFL says early access can generate video with audio up to 20 seconds in one generation.
Build Character Scenes with Native Sound
The announced model combines visual generation and audio, including multilingual dialogue, while supporting starting frames and visual references.
Multilingual Dialogue
Generate spoken dialogue in multiple languages as part of the native audio workflow.
Image-to-Video Options
Animate a starting frame or use an image as a broader visual reference for the scene.
Reference Consistency
Use approved visual inputs to help ground character appearance and scene direction.
Chain Dynamic Shots into Longer Sequences
FLUX 3 Video is intended to support continuation, keyframe transitions, and multi-shot workflows that can be assembled into a longer edit.
Video and Audio Continuation
Extend an existing input clip while continuing both its visual and audio direction.
Keyframe Transitions
Define visual endpoints and generate motion that connects those keyframes.
Multi-Shot Chaining
Join individual clips into longer sequences and use references to support character continuity.
What FLUX 3 Video Is Designed to Do
Based on Black Forest Labs' July 23, 2026 early-access announcement. Results and specifications remain preliminary while the model is in development.
Text-to-Video with Native Audio
Generate video and synchronized audio together, including dialogue and sounds tied to physical events.
Two Image-to-Video Workflows
Animate a starting frame or use an image as a visual reference without requiring it to be the first frame.
Video-to-Video
Use a source clip as a reference for motion, structure, or visual transformation.
Video and Audio Continuation
Continue an input clip while extending both the moving image and its associated audio.
Keyframes and Multi-Shot Chaining
Create transitions between defined keyframes and combine individual generations into longer sequences.
Styles, Typography, and Motion Design
The preview covers varied aesthetics, animated typography, and design-led motion workflows.
A Future FLUX 3 Video Workflow
Video generation is not available yet. When integration opens, the intended production workflow will follow three stages:
Write the Scene, Sound, and Shot Plan
Define the subject, action, setting, camera movement, timing, atmosphere, dialogue, and important sound events.
Add Approved Image or Video References
Choose a starting frame, visual reference, source video, continuation clip, or keyframes according to the intended workflow.
Generate, Review, and Finish the Edit
Review motion, continuity, audio synchronization, and character consistency before refining prompts or assembling clips into a longer sequence.
FLUX 3 Video: Availability and Details
What early access means, what has been confirmed, and what remains unavailable on Flux2Pro.
Can I generate FLUX 3 Video on Flux2Pro now?
Not yet. The generation controls are intentionally disabled, and this model page is marked Coming Soon.
Is FLUX 3 Video officially released?
Black Forest Labs opened early access on July 23, 2026. BFL also states that the model and its surrounding harness remain in development, so current evaluations and specifications are preliminary.
Does FLUX 3 Video generate audio?
Yes. BFL says all announced video workflows include native audio generation, including multilingual dialogue and sounds associated with physical events.
What inputs does it support?
BFL has announced text-to-video, two image-to-video workflows, video-to-video reference transfer, video-audio continuation, and keyframe-to-video transitions.
What is the maximum video length?
The early-access announcement states that FLUX 3 can generate video with audio up to 20 seconds in one generation. Final Flux2Pro limits have not been set.
Can it create longer multi-shot sequences?
BFL says individual clips can be chained into longer sequences, with visual references used to support character consistency across shots.
Will FLUX 3 Video be available with open weights?
BFL's plan includes FLUX 3 Dev, an open-weight multimodal backbone. Final availability, model size, hardware requirements, and license terms have not been published.
Is Flux2Pro affiliated with Black Forest Labs?
No. Flux2Pro is an independent service and is not affiliated with, endorsed by, or sponsored by Black Forest Labs.