What is FLUX 3 and who built it?

FLUX 3 is a multimodal foundation model from Black Forest Labs, the lab behind the FLUX image models. Announced on July 23, 2026, it generates video with native audio, synthesizes and edits images, and learns all modalities in one architecture — a departure from FLUX 1 and FLUX.2, which were image-only.

What can FLUX 3 do for video?

Text-to-video, image-to-video animating a start frame (with optional last frame), storyboard keyframes that turn 3–10 images into a directed sequence, and clip extension that continues existing footage. Single generations run 5–20 seconds at 720P or 1080P, always with the option of native audio — all four workflows are live in the generator on this page.

Does FLUX 3 generate audio with video?

Yes — natively. Every video output can include audio generated by the same model in the same pass: multilingual dialogue synced to lip movement and sound effects tied to on-screen physical events. There's no separate audio model or post-sync step, and turning audio on costs nothing extra. The 20-second monologue sample above was generated from a three-word prompt, voice included.

How does the storyboard keyframe mode work?

Upload 3–10 keyframe images in reference mode and FLUX 3 spreads them across the timeline, generating the motion between your exact frames — storyboard-level control over composition and pacing that pure text prompts can't reach. Reference mode is dedicated to keyframes; continuing an existing clip is a separate model, FLUX 3 Extend.

Can FLUX 3 extend an existing video?

Yes — pick the FLUX 3 Extend model, upload one clip and describe what happens next; FLUX 3 continues the footage while keeping the look and motion consistent. The newly generated seconds bill at 91 credits per second at 720P or 118 at 1080P — the source clip itself costs nothing.

What about image generation?

FLUX 3 synthesizes and edits images across a wide range of styles, aspect ratios, and resolutions, with markedly better complex-prompt handling than earlier FLUX generations and accurate multilingual text rendering. The image side is still in early access — on Molyin the image slot stays marked coming soon and activates the day access opens, with GPT Image 2 and Nano Banana Pro one click away until then.

How many credits does a FLUX 3 video cost?

FLUX 3 bills 38 credits per second at 720P and 65 at 1080P, audio included. Concretely: a 5-second 720P clip is 190 credits, 10 seconds is 380, the same 10 seconds at 1080P is 650, and a full 20-second 720P take is 760. Extending an existing clip bills 91/118 credits per second for the new footage only. Failed generations are refunded automatically.

How do I prompt FLUX 3 for dialogue and sound?

Put spoken lines in quotes and add a tone direction — FLUX 3 renders the voice, lip-synced. Pin sound effects to the physical events that cause them ("the glass shatters on the tile — a sharp crack, then skittering shards") instead of listing sounds in isolation. FLUX 3 also stages surprisingly coherent scenes from vague prompts — the monologue and pyramid samples above used a handful of words — but specific beats give you control. The full prompt guide linked above covers dialogue, effects, and storyboard patterns.

How does FLUX 3 compare to Seedance, Kling, and Runway?

In early human-preference tests of 10-second 720P clips with audio, FLUX 3 was preferred over Runway Gen-4.5 in 77% of comparisons, Kling v3 Pro in 60%, HappyHorse 1.1 in 57%, and Seedance 2.0 in 52%. Those are pre-release numbers — a strong early signal, not a settled verdict. The comparison table above shows how it positions against every model live on Molyin, and you can switch between them in the generator to judge for yourself.

Will FLUX 3 have open weights?

Yes — the launch plan includes FLUX 3 Dev, an open-weight multimodal backbone for content creation (video, audio, image) and action prediction, following the pattern of earlier FLUX open releases. Timing hasn't been announced.

Is there a FLUX 3 API?

Yes for video — FLUX 3 video is live on Molyin today: 5–20 seconds at 720P/1080P with native audio, storyboard keyframes, and clip extension. The image side is still in early access; on Molyin it stays marked coming soon and activates the day access opens.

Can I use FLUX 3 on Molyin?

Yes — FLUX 3 video is live in the generator at the top of this page, on the same credit system and workflow as every other model, no separate subscription. Remix any sample on this page to start from a proven prompt, or write your own from scratch.

FLUX 3

FLUX 3: One Multimodal Model for Video, Image & Audio

Black Forest Labs' world model generates picture and sound in a single pass — write a prompt below and FLUX 3 renders 5–20 seconds of footage with dialogue and effects built in.

Failed generations auto-refunded
Credit packs never expire
Every model's price is public
Cancel anytime in one click

5–20 s
Clip length: 720P / 1080P
Resolution: Native, on/off
Audio: 3–10
Storyboard keyframes

Real Prompts, Real FLUX 3 Footage

Every clip below was generated by FLUX 3 from the exact prompt shown — official samples and community work, sound and all. Hit Remix to load any prompt into the generator above and make it yours.

FLUX 3
Text to video
10s
1080P

A cozy ramen shop on a rainy Tokyo night: steam rising from the broth, neon reflections in the window puddles, the cook working calmly. The camera drifts slowly past the counter. Rain patter and quiet kitchen sounds.

FLUX 3
Text to video
20s
720P

ranting about ai

FLUX 3
Text to video
20s
720P

How pyramids are made? Visualize it.

FLUX 3
Text to video
15s
720P

claymation story about a man consistently getting smushed by a large giant hand that enters the scene so he cant live his life

FLUX 3
Text to video
20s
720P

Daily life of the Aztecs

One Photo In, a Fireworks Finale Out

FLUX 3's image-to-video mode animates a single still into motion. Here is the official example end to end: one fireworks photo plus a one-line prompt becomes a 10-second 1080P finale — crowd murmur and explosions generated natively with the picture.

Prompt

A fourth of July fireworks display concludes with a dazzling finale of red, white and blue fireworks.

10s · 1080P · native audio — generated by FLUX 3 image-to-video

What Is FLUX 3?

FLUX 3 is the multimodal foundation model announced by Black Forest Labs — the lab behind the FLUX image family — on July 23, 2026. Where FLUX 1 and FLUX.2 generated images, FLUX 3 learns from images, videos, and audio jointly inside one architecture. The bet is simple: a model that must render how objects move, collide, and sound is forced to learn how the world actually behaves — and every modality it generates benefits from that shared understanding.

That world-model foundation shows up in the footage. FLUX 3 generates clips up to 20 seconds long with native audio in a single pass — dialogue in multiple languages synced to lip movement, sound effects pinned to the physical events that cause them, styles ranging from handheld camcorder footage to claymation and cinematic shots. Give it three words and it stages a coherent scene; give it storyboard keyframes and it directs the transitions between your exact frames. On the image side it handles complex prompts markedly better than earlier FLUX generations and renders accurate multilingual text.

FLUX 3 video is live on Molyin today — the video side of the AI video generator above runs it directly, with storyboard keyframes and clip extension included. The image side is still in early access, so FLUX 3 stays marked coming soon in the AI image generator — the slot is ready, and both sides run on one credit system that covers every model.

FLUX 3 vs. Seedance 2.5

The two newest flagship video models on Molyin, side by side. They complement rather than replace each other: FLUX 3 leads on native dialogue and keyframe direction, Seedance 2.5 on 30-second takes and reference capacity — one credit balance covers both.

Capability	FLUX 3	Seedance 2.5
Native audio	Dialogue + sound effects in the same pass, toggle on/off at the same price	Joint audio-visual generation, always on
Clip length	5–20 s, any length in range	4–30 s, any length in range
Resolution	720P / 1080P	480P / 720P
Reference inputs	3–10 storyboard keyframes drive the shot sequence	Up to 50 reference assets — images, clips, and audio
Clip extension	FLUX 3 Extend model — continues an existing clip, 5–20 s	Seedance 2.5 Extend model — continues from the last frame, 4–30 s
Multilingual dialogue	Characters speak multiple languages, lip-synced	Multilingual lip-sync — eight languages demonstrated

Both models are live in the generator above — switch between them without leaving the page.

FLUX 3 vs models live on Molyin

Every row in this table is live on Molyin right now. Native dialogue, storyboard keyframes, 30-second takes, 4K — pick the tool per shot, one credit balance for all of them.

Model	Positioning	Release Date	On Molyin
FLUX 3	Black Forest Labs' world model — native audio + storyboard keyframes	Aug 7, 2026 (video API) · open weights announced	Live — generate above
Seedance 2.5	ByteDance flagship — 30s takes, directed second by second	Aug 7, 2026 (API)	Live — generate above
Wan 3.0	Alibaba's All-in-One model, up to 30s takes	Aug 6, 2026	Live — generate above
MiniMax H3	Omni-modal flagship, native 2K + stereo	Jul 31, 2026 · open weights announced	Live — generate above
HappyHorse	Expressive generation + instruction video editing	Jun 2026 (1.1)	Live — generate above
Seedance 2.0 family	Standard / Fast / Mini workhorses, up to 4K	Apr 2026 (API)	Live — generate above
Kling 3.0	Element references + native audio, up to 4K	Feb 4, 2026	Live — generate above
Veo 3.1	Google's cinematic model with audio	Oct 2025	Live — generate above

One subscription, one credit balance — every live model above is included.

FLUX 3 Full Specifications

Every FLUX 3 parameter on Molyin, straight from the model registry and billing engine — what this table says is what the generator enforces.

FLUX 3 Full Specifications
Parameter	Value	Notes
Generation modes	Text-to-video · Image-to-video (first/last frame) · Storyboard keyframes · Clip extension	Storyboard keyframes run in reference mode; clip extension is its own model — pick FLUX 3 Extend in video-extend mode.
Resolution	720P / 1080P	Both tiers generate with native audio; each tier has its own per-second rate.
Clip length	5–20 seconds	Any length in the range — billing is per second of output.
Aspect ratios	auto · 16:9 · 9:16 · 4:3 · 3:4 · 1:1 · 21:9 · 2:1	auto lets the model pick from the prompt and inputs; 2:1 widescreen is a FLUX 3 exclusive on Molyin.
Audio	Native — dialogue and sound effects, toggle on/off	Generated in the same pass as the picture; same price with audio on or off.
Languages	Multilingual dialogue and on-screen text	Characters speak multiple languages lip-synced, and rendered signs or titles stay legible.
Storyboard keyframes	3–10 images	Keyframes are spread across the timeline; the model generates the motion between your exact frames.
Clip extension	1 source clip	The FLUX 3 Extend model continues an existing clip; new footage bills at the extension rate below.
Seed	Not supported	Vary the wording to explore alternatives — identical prompts still differ between runs.
Credits per second	38 (720P) · 65 (1080P)	Text-to-video, image-to-video, and storyboard keyframes all bill at this rate.
Extension credits per second	91 (720P) · 118 (1080P)	Applies only to newly generated footage when extending a clip — the source clip itself is free.

Everything listed here is already supported on the site, and stays in sync with the model's latest official capabilities.

What FLUX 3 Costs in Credits

Transparent per-second pricing from the same billing engine that charges the generator — audio adds nothing, and failed generations are refunded automatically.

AI video model credit rates by resolution — credits per second and per-clip examples
Model	Resolution	Credits/sec	5s Video	5s Video + 3s Reference
FLUX 3	720p	38	190 credits	Reference images only
FLUX 3	1080p	65	325 credits	Reference images only
FLUX 3 Extend	720p	91	455 credits	455 credits (source video free)
FLUX 3 Extend	1080p	118	590 credits	590 credits (source video free)

Video-extend models (FLUX 3 Extend) never bill the source video — only the newly generated seconds are billed, at the extend rate shown above.

See credit pricing for every model

How to generate with FLUX 3

Three steps from a prompt to footage with sound

Pick FLUX 3 and a mode

Choose text-to-video, image-to-video, or reference mode for storyboard keyframes; to extend a clip, pick the FLUX 3 Extend model.

Write the scene, sound included

Describe the shot and put dialogue in quotes — FLUX 3 renders the audio you write. Attach keyframes in reference mode, or hand FLUX 3 Extend a source clip, to direct the timeline.

Generate and export

FLUX 3 delivers picture and sound in one pass, 5 to 20 seconds at 720P or 1080P. Preview in place, then download the original file.

Write sound into your prompts

FLUX 3 renders the audio you describe — pin each sound cue to the physical event that causes it and the mix follows the picture. Straight from our full FLUX 3 prompt guide.

[Visual event] with [its sound] + optional [ambient bed]

Required

Visual Event

The on-screen action that makes noise: a door slams, glass clinks, footsteps splash.

Required

Sound Cue

The effect pinned to that event — named concretely, in the same sentence as the action.

Optional

Ambience

The background bed: wind, crowd murmur, room tone. One phrase is enough.

Read the full FLUX 3 prompt guide

FLUX 3 FAQ

What creators ask about Black Forest Labs' multimodal model.

Start Creating with Molyin Today

Start Creating Free View Pricing