Black Forest Labs video model

FLUX.3 Video. Keyframes, continuation, and 20-second takes.

Black Forest Labs brings FLUX discipline to video: exact first and last frames, clips up to 20 seconds with audio, and continuation that extends footage you already have.

Try FLUX.3 Video->See all models

5 to 20 second clips
6 aspect ratios
Native audio
First and last frame control

The model

The team behind the FLUX image models built their video model around control. FLUX.3 Video generates from text or from pinned first and last keyframes, runs 5 to 20 seconds per clip, which is the longest range in the Treza lineup alongside Sora 2 Pro, and outputs 720p or 1080p with native audio across six aspect ratios up to 21:9 cinema.

Its distinctive trick is video continuation: hand it an existing clip and it keeps going, extending a take instead of cutting to a new one. On Treza that composes naturally, with one node generating an opening shot and a continuation pass stretching it into a long take, then captions and publishing downstream. Generation bills 17 cents per second at 720p and 29 at 1080p, continuation at 41 and 53.

Video continuation

Extend existing footage instead of starting over. The only model in the lineup that grows a take you already like.

Up to 20 seconds per clip

5 to 20 second range with native audio, enough for a full scene beat in one generation instead of a stitch.

Keyframes at both ends

Pin the opening and closing frame and the model interpolates the journey, which is how storyboards become footage predictably.

1080p in six aspect ratios

720p or 1080p output from 21:9 cinema to 9:16 vertical, so the same model covers hero film and Shorts.

Prompt to result

What to make with FLUX.3 Video

Real prompts, and what comes back. Every one runs in Treza chat or as a node on the canvas.

“A hot air balloon drifts over vineyards at sunrise, 20 seconds, 21:9”

A full scene beat in one uncut generation.

“Continue this clip: the drone keeps climbing until the whole coastline is visible”

A continuation pass that extends the original take.

“First frame: closed book on a desk. Last frame: the book open under lamplight”

Keyframe-to-keyframe motion matching the storyboard.

“A glassblower shapes a vase, workshop ambience, 1080p”

Craft footage with native audio at full HD.

“A storm rolls across a wheat field in one slow 15-second push-in”

A patient long take most models cannot hold.

“A metro train arrives, doors open, crowd flows out, 9:16”

Vertical documentary-style footage for short-form.

Beyond one prompt

FLUX.3 Video is one node on a canvas

Chain it with language, image, and video models, then publish the whole pipeline as an API.

Step 01
Draft the brief
A language model node expands a one-line idea into a detailed, on-brand prompt before it ever reaches the model.
Step 02
Generate with FLUX.3 Video
Run variants in parallel, compare against other models on the same prompt, and keep every result in run history.
Step 03
Publish it as an API
The pipeline becomes a versioned endpoint. Your product calls it with the typed /invoke API or any OpenAI SDK.

FAQ

FLUX.3 Video, answered

What is video continuation and why does it matter?

Continuation takes an existing clip as input and generates what happens next, preserving the scene instead of cutting away. It is how you build long takes and extended sequences: generate an opening you like, then continue it, rather than rolling new clips and hiding the seams in edits.

How long can a FLUX.3 Video clip be?

5 to 20 seconds per generation, and continuation passes extend existing footage beyond that. For assembled pieces, the Sequence node and timeline editor take it from there.

What does it cost?

17 cents per second of output at 720p and 29 cents at 1080p, with continuation at 41 and 53 cents per second. It bills from prepaid credits and only successful generations are charged.

How much does FLUX.3 Video cost on Treza?

Treza uses prepaid credits with no subscription. Each clip is charged at FLUX.3 Video's rate from your balance, and only successful generations are charged. Credit packs start at $5 and never expire.

Can I call FLUX.3 Video through an API?

Yes. Build a pipeline that uses FLUX.3 Video and publish it as a versioned HTTP endpoint. Call the typed /invoke API from your product, or point any OpenAI SDK at the OpenAI-compatible endpoint. If a better model ships later, swap it on the node without changing the API your product calls.