Pinterest

FLUX 3

FLUX 3 jointly learns images, video, and audio in one architecture, generating 20-second clips with native sound. Put it to work inside VideoGen workflows.

FLUX 3Generate

By Black Forest Labs · Announced July 2026; video generally available August 2026 · Multimodal frontier model; strong on facial expression, event-synced sound, and lip-synced multilingual dialogue

Trusted by over 5 million creators and teams

Example outputs from FLUX 3

Real prompts, real outputs, generated with FLUX 3 in VideoGen.

“A blacksmith hammers glowing steel on an anvil in a dim workshop, each strike lands with a sharp metallic ring and a burst of sparks, bellows breathe in the background”

Generated with FLUX 3 in VideoGen

A barista slides a latte across the counter and says 'This one's on the house', cozy cafe ambience with espresso machine hiss and low chatter, warm tungsten light

“Close-up of a barista finishing latte art, she looks up with a warm smile and says 'One rosetta, just for you', espresso machine hisses and cafe chatter hums behind her”

Generated with FLUX 3 in VideoGen

“A neon sign reading 'OPEN ALL NIGHT' flickers to life above a rainy diner entrance, each letter buzzes as it ignites, reflections shimmer in puddles on the sidewalk”

Generated with FLUX 3 in VideoGen

How VideoGen integrates FLUX 3

Model specifications for FLUX 3
SpecificationDetails
Max duration20 seconds per generation
ResolutionsHD (720p) and Full HD (1080p via upscaling)
Native audioYes: dialogue, sound effects, and ambient sound
ModesText-to-video, image-to-video, keyframes, video continuation, multi-shot scenes in one generation
Input typesText, image references, start and end frames, up to 4 seconds of video and audio
Languages14+ lip-synced languages, including English, Chinese, Spanish, Hindi, and Japanese
Draft modeFast low-cost previews; approved drafts re-render at full quality
Open weightsFLUX 3 Dev backbone announced for later release
Elo BenchmarkUnknown

Quality tiers in VideoGen

Low

The fastest and most affordable tier, built for drafts and quick iteration.

Standard

The balanced default: good quality video clips at an everyday cost.

High

Premium quality video clips for the content you publish.

Max

MAX always integrates the state of the art. FLUX 3 is part of that set.

Every quality tier routes each request to the best model based on our internal evals, compliance requirements, and your intent. Requests stay fully compliant at every tier: sensitive data is never routed to models hosted by foreign adversaries such as China.

How to use FLUX 3 in VideoGen

1

Create a free account

Sign up for VideoGen to open the AI playground and editor.

2

Open a workflow

Start the storyboard to video or prompt to video workflow, or use the AI video tool inside any project.

3

Pick a quality tier or type a prompt

Choose the tier that fits your needs and describe what you want to create.

4

Generate and build your video

VideoGen builds the finished video with visuals, narration, and captions, ready to export.

FLUX 3 API

Developers can generate video clips through the VideoGen API with the same quality tier routing as the app, so requests reach models like FLUX 3 without managing provider keys.

curl -X POST "https://api.videogen.io/v1/tools/generate-video-clip" \
  -H "Authorization: Bearer $VIDEOGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "A drone shot over a coastal town at sunrise", "durationSeconds": 8}'

Learn more about the VideoGen API Read the API docs

One model, every visual

Generate video with synchronized sound and high-fidelity images from the same multimodal backbone, keeping style consistent across your assets.

“VideoGen is awesome for scaling video production and improving turnaround time... The complex process of video editing, which could take days or months, now takes minutes!”

Ary Aranguiz
Ary Aranguiz
Learning Manager, Google

“VideoGen solves the biggest pain points of video production—complexity, cost, and time. In just a few clicks, anyone can create professional, copyright-free videos.”

Garry Tan
Garry Tan
CEO of Y Combinator

“VideoGen is the most underrated tool for content creators who want to put out the highest quality content in the shortest amount of time.”

Andrew Yu
Andrew Yu
6M+ Followers

Frequently asked questions

Go from FLUX 3 outputs to finished videos

Every VideoGen workflow uses models like FLUX 3 automatically to build complete videos with visuals, narration, music, and captions.

Start creating with AI on VideoGen.

Sign in to generate images, video clips, and more with the models you choose.

Sign in to generate

Related AI video models

View all AI video models...