Official launch media01
A world that moves
FLUX 3 / BFL
Independent early-access preview
Imagine it.
Make it move.
Explore image, video and native-audio workflows for FLUX 3 — one cinematic playground built for ideas that need to look, move and sound alive.
01 Image02 Video03 Native audio04 Action
OFFICIAL BFL MEDIA00:20 ↗
One model.
Every modality.▶
Every modality.▶
01 / THE OMNI WORKFLOW
Stop stitching separate tools together. Shape the whole scene in one creative flow.
FLUX 3 is positioned as more than an image model: a jointly trained system for visuals, motion, sound and action. FLUX/3 RUN turns that direction into a clear, creator-first workspace.
02 / PLAYGROUND
From prompt to moving world.
Build an image or video brief, add references, then preview the intended output workflow.
03 / MOVING GALLERY
Made to be watched,
not just scrolled past.
A media-first view of the image-to-video, native-audio and scene-continuation workflows FLUX 3 is built to explore.
Illustrative demo02
Bioluminescent response
Video + sound concept
Official launch film03
One multimodal model
Image / video / audio / action
Reference input04
Continue an existing scene
Video-to-video workflow
Official launch media is labeled “Official.” Other clips are illustrative workflow references and are not represented as official FLUX 3 generations.
04 / CAPABILITIES
One model.
Four ways to create.
Start with language, a frame or a moving clip. Direct the result as one coherent world instead of four disconnected exports.
Image generation
Create expressive, controlled stills from text and multimodal references.
↗Video generation
Move from a single idea or image into cinematic motion and camera behavior.
↗Native audio
Conceptualize dialogue, ambience, effects and music together with the visual scene.
↗Action prediction
Explore world models that understand what happens next, not only what a frame looks like.
↗Reference & continue
Preserve subjects, environments and visual language while extending existing media.
↗Scene continuity
Keep the creative thread coherent across movement, sound and iterative refinements.
↗05 / HOW IT WORKS
Direct the world,
frame by frame.
Describe what should happen
Write the subject, setting, camera, movement, light and sound in natural language.
Add visual memory
Attach images or a short reference clip to preserve identity, composition or motion language.
Choose the output
Switch between still image and video, then set aspect, duration and resolution.
Generate, continue, refine
Preview the workflow today. Live output activates when authorized FLUX 3 access is available.
EARLY ACCESS
Be first to
run the future.
Get notified when live FLUX 3 image and video generation becomes available on FLUX/3 RUN.
06 / FAQ
Before you
press run.
Clear answers about model status, media sources and what this independent preview can do today.
What is FLUX 3?+
FLUX 3 is presented by Black Forest Labs as one multimodal model spanning image, video, audio and action prediction. BFL currently labels it Coming Soon and offers an early-access request.
Can FLUX 3 generate video with audio?+
BFL describes native video and audio as part of the FLUX 3 multimodal direction. Public availability, limits and final API behavior may change during rollout.
Is FLUX/3 RUN an official Black Forest Labs website?+
No. FLUX/3 RUN is an independent preview experience and is not affiliated with, endorsed by or operated by Black Forest Labs.
Can I use a reference image or video?+
The playground is designed around text, reference images and reference video workflows. During Early Access, demo previews illustrate the intended control surface before an authorized generation provider is connected.
When will generation be available?+
We will activate live generation when an authorized FLUX 3 API or approved provider becomes available. The site does not claim public model access before that point.