Black Forest Labs video model with native sound
Flux 3 Video AI Generator
Flux 3 Video makes the picture and the sound in one pass, so every hammer strike rings on the hit and a line you put in quotes comes back spoken word for word.
Flux 3 Video is Black Forest Labs' text to video and image to video model. On Pixel Dojo it makes 5 to 20 s clips at 720p or 1080p in six aspect ratios, from 21:9 widescreen to 9:16 vertical, with sound effects and dialogue built in. We made every clip on this page on Pixel Dojo, and each one comes with the exact prompt so you can start from it.
25 credits ($1.56) per 5 s clip at 720p with audio. Included in every plan, from $10/month.
Already a member? Sign in
By Brian Lovett · Updated
What you can create with Flux 3 Video
We made these four clips on Pixel Dojo, each in a different genre and all with sound on: a forge scene, a spoken line, a storm coast aerial and a hand-drawn anime reveal. Tap Use this prompt on any clip to start from our exact wording.
A 10 s, 1080p clip in one continuous take. The camera pushes in slowly as sparks jump off the anvil, and every ringing strike lands the moment the hammer meets the steel. Start from this prompt for any craft, maker or workshop scene.
Low-angle medium shot, slow push-in, one continuous take. A grey-bearded blacksmith in a scorched leather apron, sleeves rolled over soot-streaked forearms, strikes a glowing orange bar of steel on an anvil, and every hammer blow throws a burst of bright sparks across the dark workshop. Coal forge glowing behind him, hard warm firelight on his face, deep black shadows, 35mm lens, fine film grain. Audio: ringing hammer strikes in time with each blow, crackling coals, a short hiss of steam, no music.
Dialogue with native sound. We put her line in quotes and she speaks it word for word with a tired smile, rain on the wheelhouse glass and the gauges lit beside the wheel. Start from this prompt for any character who talks to camera.
Medium close-up, slow push-in, one continuous take. A ferry captain in her fifties with wind-tangled grey hair, a navy knit cap and a weathered yellow oilskin jacket grips the wheel of a small boat in the rain, glances toward the camera and says with a tired smile, "Hold on tight, it gets rough past the point." Rain streaking the wheelhouse window, cold blue dusk light and a warm amber instrument glow on her face, 50mm lens, soft film grain. Audio: her voice clear and close, rain drumming on the roof, a creaking hull, low engine rumble, no music.
A native 21:9 ultrawide aerial. The drone glides low over dark green swell, a shaft of light breaks through the clouds, and a hiker in a yellow raincoat stands tiny on the basalt cliff as the waves burst into white spray. Start from this prompt for travel and adventure openers.
Ultrawide aerial, slow forward drone glide low over the water, one continuous take. Huge dark green waves roll toward a black basalt sea cliff and explode into towering white spray, while a lone hiker in a bright yellow raincoat stands on the clifftop, tiny against the storm. Overcast silver light breaking through one gap in the clouds, wind-blown mist, anamorphic lens, muted teal and slate grade with a pop of yellow. Audio: deep booming surf, roaring wind, distant gulls, no music.
Hand-drawn anime with sound on. A moss-covered stone guardian rises from a mountain lake, amber eyes glowing and water pouring off its shoulders, as the white herons on its back take flight past painted peaks. Start from this prompt for fantasy and animated shorts.
Wide shot, slow crane up, hand-drawn 2D anime style with lush painted backgrounds. A giant stone guardian with a mossy boulder head and two glowing amber eyes rises out of a still mountain lake, its head and shoulders already breaking the surface, sheets of water pouring off its shoulders as a flock of white herons bursts from its back into the morning sky, pine forest and snowy peaks behind. Soft golden morning light, visible line art, flat cel shading, vivid greens and sky blues. Audio: rushing water, a deep stone rumble, beating wings, one soft flute note.
How to make a Flux 3 Video clip on Pixel Dojo
- 1
Describe the shot like a camera note
Open with shot size and one camera move, then one subject doing one clear action, then light and lens. Our forge prompt starts "Low-angle medium shot, slow push-in, one continuous take," and the clip follows that move for the full 10 s.
- 2
Write the sound and the dialogue
End the prompt with an Audio: line that names two or three sounds, and put any spoken words in quotes. We asked our ferry captain to say "Hold on tight, it gets rough past the point," and she said it word for word.
- 3
Pick the shape, length and resolution
Choose one of six aspect ratios, a length from 5 to 20 s, and 720p or 1080p. A 5 s clip at 720p is 25 credits, and our 10 s forge clip at 1080p was 90 credits. Attach a still to animate it, or a clip to extend it.
Where Flux 3 Video shines, with tips from our clips
We made every clip on this page with the same Flux 3 Video tool you get on Pixel Dojo. Here is what it does best, plus one more genre: a vertical product shot made for social feeds.
A 9:16 vertical product shot. Cream pours into a ridged glass of iced coffee and curls through it in slow marbled clouds, with beads of water on the black slate. Start from this prompt for drink, cafe and beverage ads.
Vertical macro product shot, slow push-in, one continuous take. A thin ribbon of white cream pours into a tall ridged glass of iced black coffee on wet black slate, blooming into slow marbled clouds that curl around the clear ice cubes as they crack and shift, while a thin band of warm light sweeps across the glass. Black studio background, glossy reflections, beads of condensation, 100mm macro lens, shallow depth of field. Audio: cream pouring, ice cracking and clinking, a soft low hum, no music.
- Sound that lands on the action. In both of our forge clips the ringing strikes arrive the moment the hammer meets the steel.
- Dialogue straight from your prompt. Put a line in quotes and Flux 3 Video speaks it; our ferry captain delivers hers word for word with a warm, tired smile.
- Long single takes. Our forge clip holds one slow push-in for the full 10 s at 1080p with no cut. For longer clips, give the subject an action that repeats, like a hammer strike.
- Six shapes for every screen. Our coast aerial ran at 21:9 and the cold brew ad at 9:16, and each one is framed for its shape.
- Live action or animation. The same model gave us photoreal firelight at the forge and hand-drawn cel shading with painted backgrounds for the guardian.
- Name the details you want to see. Our guardian prompt spells out a mossy boulder head and two glowing amber eyes, and both show up as it rises from the lake.
Flux 3 Video vs WAN 3.0 on one blacksmith prompt
We gave both models the same forge prompt at 720p, 5 s, 16:9 with sound on. WAN 3.0 went bright and fiery for 9 credits. Flux 3 Video stayed true to the dim, shadowy workshop in the prompt, kept the steel bar straight through every strike and timed each ring to the hammer. Pick Flux 3 Video when sound and continuity carry the shot.
Same prompt: Low-angle medium shot, slow push-in, one continuous take. A grey-bearded blacksmith in a scorched leather apron, sleeves rolled over soot-streaked forearms, strikes a glowing orange bar of steel on an anvil, and every hammer blow throws a burst of bright sparks across the dark workshop. Coal forge glowing behind him, hard warm firelight on his face, deep black shadows, 35mm lens, fine film grain. Audio: ringing hammer strikes in time with each blow, crackling coals, a short hiss of steam, no music.
Flux 3 Video. Natural firelight and deep shadows as the prompt asked, the glowing bar holds its shape across the anvil, and each strike sound lands as the hammer meets the steel.
Low-angle medium shot, slow push-in, one continuous take. A grey-bearded blacksmith in a scorched leather apron, sleeves rolled over soot-streaked forearms, strikes a glowing orange bar of steel on an anvil, and every hammer blow throws a burst of bright sparks across the dark workshop. Coal forge glowing behind him, hard warm firelight on his face, deep black shadows, 35mm lens, fine film grain. Audio: ringing hammer strikes in time with each blow, crackling coals, a short hiss of steam, no music.
WAN 3.0 on the identical prompt. A brighter, higher contrast look with a roaring forge fire behind him and a big burst of sparks on the strike.
Low-angle medium shot, slow push-in, one continuous take. A grey-bearded blacksmith in a scorched leather apron, sleeves rolled over soot-streaked forearms, strikes a glowing orange bar of steel on an anvil, and every hammer blow throws a burst of bright sparks across the dark workshop. Coal forge glowing behind him, hard warm firelight on his face, deep black shadows, 35mm lens, fine film grain. Audio: ringing hammer strikes in time with each blow, crackling coals, a short hiss of steam, no music.
| Flux 3 Video | WAN 3.0 | |
|---|---|---|
| Credits | 25 credits | 9 credits |
| Look | Dim firelight and deep shadows, true to the prompt | Bright, high contrast, big open fire |
| Steel bar | Holds its straight shape through every strike | Glowing metal with bold spark bursts |
| Sound | Strike sounds timed to each hammer hit | Stereo sound track |
| Best for | Scenes where sound timing and object shape carry the shot | Bright, punchy frames at a lower credit cost |
Flux 3 Video settings and what each one does
The tool has one prompt box. Leave the media slots empty for text to video, attach a still for image to video, or attach a clip to extend it. Black Forest Labs also built Flux 3 Video for multi-shot sequences and on-screen typography, both driven by your prompt.
- Resolution720p · 1080p
- 720p is 5 credits per second and 1080p is 9 credits per second. Our 10 s forge clip at 1080p is full HD, ready to cut into an edit.
- Duration5 to 20 s
- Give longer clips an action that repeats or builds. Our forge prompt repeats the hammer strike, and it carries the full 10 s in one take. Extensions of a clip you already have run up to 15 s.
- Aspect ratio21:9 · 16:9 · 4:3 · 1:1 · 3:4 · 9:16
- Set it for text to video. Our coast aerial ran 21:9 and the cold brew ad ran 9:16, each framed for its screen. With a start frame or a source clip, the upload sets the shape.
- Generate audioOn · Off
- On by default, at the same price either way. End the prompt with an Audio: line naming two or three sounds, and put dialogue in quotes. Switch it off when the clip will sit under your own music.
- Start and end frameImage URLs, optional
- A start frame animates any still at the text to video price. Add an end frame and the model builds the move between the two, which suits product reveals and transitions.
- Source videoVideo URL, optional
- Continue or transform a clip you already have, up to 15 s, at 13 credits per second at 720p or 16 credits per second at 1080p. Each job takes either frame images or a source clip.
Flux 3 Video pricing per clip
Flux 3 Video is included in every Pixel Dojo plan, and plans start at $10/month. Clips are priced by length and resolution, and sound comes at no extra cost.
Text to video and image to video run 5 credits per second at 720p and 9 credits per second at 1080p. A 5 s clip at 720p, like our captain, coast, guardian and cold brew clips, is 25 credits. Our 10 s forge clip at 1080p is 90 credits.
Extending a clip you already have costs 13 credits per second at 720p and 16 credits per second at 1080p. Your credits work on every other video and image model on Pixel Dojo too, so you can make a still with an image model and bring it to life with Flux 3 Video.
Flux 3 Video questions
What is Flux 3 Video?
Flux 3 Video is Black Forest Labs' AI video model. It turns a text prompt or a still image into a 5 to 20 s clip at 720p or 1080p, can extend a clip you already have, and generates sound and dialogue in the same pass. You can use it on Pixel Dojo in the Flux 3 Video tool.
What is Flux 3 Video best at?
Scenes where sound matters. It times sound effects to the action, speaks quoted dialogue and holds long continuous takes. It also covers a wide range of looks: our examples include a blacksmith forge, a ferry captain talking to camera, a 21:9 storm coast aerial, a hand-drawn anime guardian and a 9:16 cold brew ad.
Does Flux 3 Video make sound and dialogue?
Yes. Sound is generated in the same pass as the picture. In our forge clips the strike sounds land on the hammer hits, and our ferry captain speaks her quoted line word for word. Black Forest Labs says it lip-syncs dialogue in more than a dozen languages. Name the sounds you want in an Audio: line at the end of the prompt.
What resolution, length and aspect ratios does Flux 3 Video support?
720p or 1080p, 5 to 20 s per clip, and six aspect ratios: 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. Extensions run up to 15 s. Every clip plays at 24 fps, and 1080p delivers full HD.
How do I use Flux 3 Video on Pixel Dojo?
Open the Flux 3 Video tool, write a prompt that starts with the shot and ends with an Audio: line, then pick length, resolution and aspect ratio. Attach a still to animate it or a clip to extend it. Developers can run the same model through the Pixel Dojo API.
How much does Flux 3 Video cost?
It is included in every Pixel Dojo plan, from $10/month. A clip costs 5 credits per second at 720p or 9 credits per second at 1080p, so a 5 s clip at 720p is 25 credits and a 10 s clip at 1080p is 90 credits. Sound is included.
Flux 3 Video vs WAN 3.0: which should I use?
Pick Flux 3 Video when sound timing and continuity carry the shot: on our forge prompt it kept the steel bar straight and put each strike sound on the hit. WAN 3.0 gave a brighter, higher contrast frame for 9 credits against 25 credits, which suits bold, punchy looks.
Make your first Flux 3 Video clip
Write a prompt above or start from any example on this page. One subscription covers every model on Pixel Dojo.
Start with Flux 3 Video