Flux 3 is Black Forest Labs' unified model — generate up to 20 seconds of flux video with native audio, plus photoreal flux image stills, from a single prompt. Start from text, an image, or a keyframe and move fluidly between formats. Free to start.
0/5000 characters
Select Model




Livestream scene — a presenter selling skincare, chat and gift animations popping up on the left. Vertical 9:16 flux video with native audio.
No Results Yet
Enter a prompt and click Generate to create your first Flux 3 result.
Native Audio
Up to 20s Video
Image + Video
Free to Start
A Flux 3 clip shared by the community.
Turn a prompt, a still image, or a keyframe into flux video up to 20 seconds long. Flux 3 handles camera motion, lighting, and physics automatically — describe the shot and it directs it.
Flux 3 generates sound as a first-class signal, not an afterthought. Dialogue, sound effects, music, and ambience are timed to on-screen events, so a closing door or a footstep lands exactly when it should.
The same model renders flux image stills across a huge range of styles — from candid camcorder footage to polished cinematics — with high-accuracy typography in multiple languages.
Give a character lines in multiple languages and Flux 3 speaks them with matching motion. Reference images keep the same character and style consistent as you chain clips into longer scenes.
Start from text, promote a still to motion, or continue an existing clip. Flux 3 lets you flow between flux image, flux video, and audio in one workflow instead of exporting between separate tools.
Go from a prompt to a Flux 3 video with sound in three simple steps.
Tell Flux 3 what you want — subject, action, mood, camera. Or upload a reference image or keyframe to guide the look, the character, or the starting frame of your flux video.
Choose a still flux image or a video up to 20 seconds with native audio. Set aspect ratio and style, hit generate, and Flux 3 handles cameras, lighting, and sound for you.
Extend the clip, swap the audio, or chain another shot while keeping your character consistent. When it's ready, download your Flux 3 result in high quality for immediate use.
Use cases
Flux 3 is built for creators and teams who need image, video, and audio to come out of one model, in sync and on brand.
Turn a prompt into scroll-stopping flux video with native audio, or a flux image thumbnail, in minutes. Keep a recurring character consistent across posts and chain 20-second clips into longer stories — no separate editing suite required.
Produce product videos, ad spots, and campaign stills from one model. Flux 3 renders accurate on-pack text, consistent lighting, and spoken multilingual voiceover, so you can localize a spot without a studio shoot.
Previz a scene, generate a keyframe, then animate it into a 20-second flux video with matched sound design. Flux 3's Self-Flow backbone keeps motion physically plausible and characters on-model across shots.
Build Flux 3 image and video generation into your own apps and pipelines through the provider API. This site is an independent web interface for Flux 3 and does not resell raw API access.
Flux 3 is Black Forest Labs' next-generation multimodal model. Instead of separate tools for image, video, and sound, Flux 3 learns them together in one Self-Flow backbone — so a still flux image, a 20-second flux video, and its native audio all come from the same model, natively in sync.
Type a scene and Flux 3 renders it as an image or animates it into flux video with dialogue, sound effects, and music. Start from text, a reference image, or a keyframe, keep characters consistent across shots, and chain clips into longer sequences.
Released
July 23, 2026
Architecture
Self-Flow (flow matching)
Max video length
20 seconds
Native audio
Dialogue, SFX & music
Inputs
Text · image · keyframe · clip
Modalities
Image + video + audio
How does Flux 3 compare to the leading AI video generators in 2026? Here is a feature-by-feature look at Flux 3 against Kling, Runway, and Luma for real creative workflows.
| Feature | Flux 3 | Kling v3 Pro | Runway Gen-4.5 | Luma Ray 3.2 |
|---|---|---|---|---|
| Native Audio | Native, synced | Partial | Post / add-on | Weak |
| Max Video Length | Up to 20s | ~10–15s | ~10–15s | ~5–15s |
| Image + Video in One Model | Yes — one model | No | No | No |
| Multilingual Dialogue | Yes | Partial | Limited | Limited |
| Character Consistency | High (references) | Good | High | Moderate |
| Keyframe / Reference Input | Text · image · keyframe | Image | Image | Keyframes |
| Style Breadth | Very high | Realistic | Cinematic | Cinematic |
| Preferred in Early Tests | Leader | 60% preferred | 77% preferred | 93% preferred |
| API Access | Provider API | API | API | API |
| Free Tier | Yes | Limited | Limited | Limited |
Flux 3 leads on unified image-plus-video generation and native, physically-aligned audio from a single model. In Black Forest Labs' early preference tests, Flux 3 was preferred over Runway Gen-4.5 and Luma Ray 3.2 by wide margins and over Kling v3 Pro in a majority of comparisons. Kling excels at human motion and Runway at post-production control, but for one model that outputs flux image, flux video, and sound together, Flux 3 is the strongest choice.
A community comparison of Flux 3 against other AI video models.
Perfect for individuals and light users
Billed annually at $90.00
Save 50%INCLUDES
For professional creators and teams
Billed annually at $234.00
Save 50%INCLUDES
Designed for large enterprises and professional studios
Billed annually at $960.00
Save 50%INCLUDES
To cancel your subscription, please contact support: support@flux-ai.video
Everything you need to know about Black Forest Labs' Flux 3 model, how to use it, and what makes it different.
Join the creators, marketers, and filmmakers already using Flux 3 to generate video with native audio and photoreal images from one model. Try Flux 3 free — sign in for free credits, no credit card required.