All Models
FLUX.3 Video
FLUX.3 generates, animates, interpolates, and extends video from text or an image.
Call the API endpoint to get the latest, detailed pricing and pricing estimates.
Call the API endpoint to get the latest, detailed pricing and pricing estimates.
Call the API endpoint to get the latest, detailed pricing and pricing estimates.
FLUX.3 Video
FLUX.3 is Black Forest Labs’ video model, released in July 2026. The Freiburg lab is known for FLUX stills; this is the video counterpart. You can generate from a prompt, animate a still, interpolate between two stills, or extend a clip. Audio is generated with the picture and is on by default.
Clips run from 5 to 20 seconds at 720p or 1080p. Aspect ratios include 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, and 9:16. Generate defaults to 9:16.
Interpolation takes a start image and an end image and invents the motion between them. Extend continues from an input video.
What it's good for
Keyframe-style work: a prompt, a start frame, a start-and-end pair, or a clip to continue. Vertical clips with sound are the default. There is no large still-reference pack. If the two stills cannot plausibly connect, interpolation will not invent a convincing path between them.
Model ID
black-forest-labs/flux-3
Output Type
Video
Pricing
From $0.17/s
Provider
Black Forest Labs
Hosting
Protoface
FLUX.3 Video
FLUX.3 is Black Forest Labs’ video model, released in July 2026. The Freiburg lab is known for FLUX stills; this is the video counterpart. You can generate from a prompt, animate a still, interpolate between two stills, or extend a clip. Audio is generated with the picture and is on by default.
Clips run from 5 to 20 seconds at 720p or 1080p. Aspect ratios include 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, and 9:16. Generate defaults to 9:16.
Interpolation takes a start image and an end image and invents the motion between them. Extend continues from an input video.
What it's good for
Keyframe-style work: a prompt, a start frame, a start-and-end pair, or a clip to continue. Vertical clips with sound are the default. There is no large still-reference pack. If the two stills cannot plausibly connect, interpolation will not invent a convincing path between them.