Prompt, Image, or Start Clip
Generate from text, lock the first frame with an image, or continue from a start video.
FLUX 3 is Black Forest Labs' multimodal video model for text-to-video, image-driven clips, and audio-synced output at 720p or 1080p.
Describe the motion, subject, camera, and mood. Add a reference when you want tighter visual control.
Move from a single generated image or video into comics, manga, storyboards, and connected visual stories.
Funktionen
FLUX 3 is built for multimodal video: prompt, images, optional start video, and audio in one model.
Generate from text, lock the first frame with an image, or continue from a start video.
FLUX 3 can generate synchronized audio with the clip. Turn it off when you need a silent preview.
Run 5 to 20 second clips at 720p or 1080p, with common cinematic and social aspect ratios.
Beispiele erkunden
Prompt-driven cinematic motion example
Reference-guided video framing example
Wie es funktioniert
Generate short videos from a prompt, a first-frame image, or a start clip.
Describe action, camera, lighting, and sound. Be specific about what should stay still.
Optionally upload a first-frame image. FLUX 3 can also continue from a start clip.
Choose duration and resolution, then generate. Use a shorter 720p pass before a 1080p final.
Paid plans only. Iterate at 720p before a 1080p final.
FAQ
No. FLUX 3 on Replicate generates video with optional audio. LlamaGen hosts it as a video model page under /m/flux-3.
Yes. One image becomes the first frame. Additional images can act as a storyboard when the backend sends them as a sequence.
FLUX 3 is available on paid LlamaGen plans. Credits scale with duration and resolution.
Gemeinschaft
Trusted Creative Infrastructure
Turn prompts and references into audio-synced clips for concept tests, ads, and cinematic drafts.
Available on paid LlamaGen plans
Creations Generated Daily
Average Creator Rating
Export Ready Quality