# Seedance 2.5 AI video generator

URL: https://dreambench.ai/models/seedance-2.5

Seedance 2.5 is ByteDance’s video model for clips of up to 30 seconds with sound, which can cut between linked shots inside one clip. On DreamBench, send it a prompt as one clip on the Video page, or let it make every scene of a planned film from your words and images, and redo any scene.

**What Seedance 2.5 is.** Seedance 2.5 is a video model from ByteDance’s Seed team, released on July 31, 2026, as the successor to Seedance 2.0. It generates picture and sound together in clips of 4 to 30 seconds, twice the 15 of Seedance 2.0, from a text prompt, images, video or audio. ByteDance built it to tell a short story in one generation: a clip can cut between linked shots, from a set-up through a turn to its ending, with the main subject keeping its look across the cuts. ByteDance says it reworked how the model renders texture, skin, eyes, light and color, so the footage looks closer to live action. BytePlus, ByteDance’s cloud, lists 11 languages for prompts and speech: Chinese, English, Spanish, Indonesian, Malay, Thai, Arabic, Portuguese, Vietnamese, Japanese and Korean. In ByteDance’s own tools it also takes up to 30 images, 10 video clips and 10 audio clips as references, edits a video at a given second and extends one. ByteDance is candid about its limits: complex motion can look physically off, and scenes where several subjects interact are not yet fully stable. On DreamBench you can use Seedance 2.5 in 2 ways. On the Video page, your prompt goes to the model as you wrote it and comes back as one clip of up to 30 seconds. In Film Studio, describe the video you want: DreamBench writes the scene plan, designs the cast or builds it from your reference images, has Seedance 2.5 make each scene and joins them into one video. If a scene misses, give it new direction and regenerate just that scene.

**What Seedance 2.5 is good at.**
- Up to 30 seconds in one clip: One generation can run from 4 to 30 seconds, long enough for a full exchange of dialogue or a scene that opens, builds and lands without a join.
- Shots that cut inside a clip: Number the shots and give each its seconds, and Seedance 2.5 cuts between them in one clip, from a wide opening to a close-up, keeping the subject’s look across the cuts.
- Sound and speech in 11 languages: Dialogue, effects, ambience and music are generated with the picture, and ByteDance lists 11 languages for speech, from English and Chinese to Indonesian, Portuguese and Arabic.
- One clip, or a whole film: On DreamBench, send a prompt to Seedance 2.5 as one clip on the Video page, or have it shoot every scene of a planned film from your references, and redo any scene.

**Seedance 2.5 on DreamBench.**
- Seedance 2.5 renders every clip at 480p, 720p or 1080p on DreamBench.
- 480p is the lowest-priced way to try an idea or a timing. When a take works, generate it again at 720p or 1080p: that is a new take, so the idea carries over but the details can differ. ByteDance’s own API has a draft mode that turns a 480p preview into its final video; DreamBench doesn’t use it.
- Sound is your choice, and on Seedance 2.5 a second of video costs the same with it on or off.
- On DreamBench’s Video page, Seedance 2.5 makes one clip of 4 to 30 seconds from your prompt exactly as you wrote it: no scene plan, no character sheets, no opening frames. It takes words only, so write the shots, their seconds and the sound yourself. To change the clip, edit the prompt and generate again.
- In Film Studio, DreamBench cuts a new film into clips of 3 to 12 seconds, each as long as its action takes, and Seedance 2.5 may cut between shots inside a clip as well. A longer video is more clips, joined: up to 30 seconds on Starter, 2 minutes on Plus and 5 minutes on Ultra and Studio.
- Add up to 8 reference images to a film, each tagged as a character, creature, location or prop, and name one in your prompt with @ and its label. DreamBench builds a sheet from each and sends it with every scene it appears in.
- Seedance won’t accept a reference image that shows a real person’s face. Describe the person instead or let DreamBench design the character: if Seedance flags a character DreamBench designed from your words, DreamBench registers it with ByteDance’s asset library and tries again. For a video of a real person, choose another model.
- To change a scene of a film, give it new direction or a new reference image and regenerate just that scene. A redo costs the seconds of the scenes you redo, plus a base amount, and the rest of the video stays as it was. A film can also stop after its script and again after its cast, so you approve them first.
- ByteDance’s own tools also take video and audio as references, use your picture as a first or last frame, edit a clip at a given second, extend it and export it as a .mov file. On DreamBench, Seedance 2.5 works from your words and reference images.

**Specifications.**
- Made by: ByteDance Seed
- Longest clip: 30 seconds per generation
- Video length on DreamBench: Up to 30 seconds on Starter, 2 minutes on Plus and 5 minutes on Ultra and Studio
- Resolution: 480p, 720p and 1080p
- Aspect ratios: 1:1, 16:9, 9:16, 4:3 and 21:9
- Sound: On or off, your choice
- Reference images: Up to 8 per video; the first 4 are included, then 43 credits each
- Plans: Every plan
- Credits: 31 per second at 480p, 57 per second at 720p and 127 per second at 1080p, plus 183 per video
- A 15-second 480p video with sound uses 648 credits.

**Prompt ideas.**
- Three shots at the harbor: Widescreen 16:9, one 12-second clip in 3 linked shots, the same fisherman, harbor and dawn light in every shot. Shot 1 (0–4 s): wide shot of a small stone harbor at dawn, a blue fishing boat chugging in through low mist, gulls wheeling overhead. Shot 2 (4–8 s): medium shot on deck as an old fisherman in a yellow oilskin coils a rope and throws it to the quay. Shot 3 (8–12 s): close-up of his weathered hands knotting the rope round an iron bollard, then of his face breaking into a tired smile. Soft pink light on the horizon, cool blue shadows, natural film grain. The diesel engine, water lapping at the hull, gulls crying, the rope creaking. No music.
- Perfume in the light: Square 1:1 product film, one 10-second clip in 2 linked shots. Shot 1 (0–5 s): a faceted glass perfume bottle of amber liquid stands on wet black stone; a single drop runs down the glass as the camera circles it slowly at label height. Shot 2 (5–10 s): macro close-up as a hand lifts the gold cap and a fine mist sprays into a beam of light, then the camera eases back to the whole bottle. Hard rim light from behind, deep shadows, glossy reflections, a polished advertising look. The soft click of the cap, the hiss of the spray, a slow, low synth melody.
- Rooftop routine at sunset: Vertical 9:16, one continuous 12-second take with no cuts, on a city rooftop at sunset. A dancer in a loose white shirt and black pants stands still, then moves into a smooth hip-hop routine: slow body rolls, a single spin at 4 seconds, sharp arm hits on the beat and, at 10 seconds, a freeze pointing at the camera. The camera stays at chest height, drifting in a slow half circle around her. Warm orange backlight, long shadows, the skyline soft behind her. A punchy hip-hop beat, sneakers squeaking on concrete, the city humming below.
- Satay at the night stall: Widescreen 16:9, one 10-second clip in 2 linked shots at a night food stall in Jakarta, the same 2 friends and warm light in both. Shot 1 (0–5 s): medium shot as a young woman in a batik shirt slides a plate of satay across the counter and says warmly, in Indonesian: “Ini sate paling enak di kota.” Shot 2 (5–10 s): close-up of her friend, a young man in a denim jacket, who takes a bite, laughs and replies in Indonesian: “Kamu benar!” Strings of bulbs, smoke from the grill drifting through the light. Charcoal crackling, scooters passing, chatter from the street. No subtitles, no music.
- The fox’s dive: Cinematic 21:9, one continuous 12-second take with no cuts, in the style of a nature documentary. In a snowy birch forest at first light, a red fox pads through deep snow, stops and tilts its head to listen, then at 8 seconds leaps and dives nose first into the drift, coming up with snow on its muzzle. The camera holds a low, steady telephoto shot from a distance, panning gently with the fox. Pale gold sun through the trees, blue shadows on the snow, the fox’s breath steaming. Snow crunching, wind in the branches, a crow calling far off. No music.
- A bowl of ramen: Classic 4:3, one 10-second clip in 2 linked shots at a small ramen bar. Shot 1 (0–5 s): overhead shot as a cook ladles steaming broth over noodles, lays slices of pork, half a soft-boiled egg and green onions on top, and slides the bowl forward. Shot 2 (5–10 s): close-up at bowl height as chopsticks lift the noodles, glistening and dripping, steam curling up into the light. Warm tungsten light, a dark wooden counter, rich, saturated color. The broth pouring, the bowl scraping on wood, a slurp, the bar’s quiet chatter.
- The red coat: Vertical 9:16 fashion film, one 10-second clip in 3 linked shots of the same model, coat and street. Shot 1 (0–3 s): wide shot as she walks toward the camera down a wet cobbled street at night in a long red wool coat. Shot 2 (3–7 s): tracking medium shot from the side, the coat swinging with each stride, her heels splashing through a puddle. Shot 3 (7–10 s): close-up as she stops, turns up her collar and looks straight into the lens. Neon reflections, rain lit from behind, glossy high-contrast color. Heels on cobbles, rain pattering, a slow, moody bass line.
- Summer platform, anime style: Widescreen 16:9 in a Japanese hand-drawn anime style, one 12-second clip in 3 linked shots, the same girl and summer light in every shot. Shot 1 (0–4 s): wide shot of a country train platform beside green rice fields, heat shimmering over the tracks. Shot 2 (4–8 s): medium shot of a schoolgirl with a short black bob holding her straw hat down as the wind gusts, watching a little train pull in. Shot 3 (8–12 s): close-up as she smiles and says softly, in Japanese: “やっと来た。” Clean line art, soft watercolor backgrounds, towering white clouds. Cicadas buzzing, the train’s bell, wind through the rice. No subtitles.
- Beach selfie for your feed: Vertical 9:16 for a phone screen, one continuous 8-second selfie take with no cuts. A young man on a sunny beach in Rio holds the phone at arm’s length, grins and says, in Brazilian Portuguese: “Bom dia! Olha só essa vista.” He turns the camera to the sea and the mountains, then back to himself. Handheld at arm’s length, gently bobbing, a natural phone-camera look. He wears a green T-shirt, sunglasses pushed up on his head. Bright morning sun, turquoise water. His voice close to the microphone, waves breaking, beach vendors calling in the distance. No subtitles, no music.
- Thirty seconds, one take: Widescreen 16:9, one continuous 30-second take with no cuts, timed in seconds. A courier on a bicycle rides through a snowy market at dusk. 0–8 s: the camera rides alongside her as she weaves between stalls hung with lights. 8–16 s: she brakes, the back wheel skidding, and stops at a bakery door. 16–24 s: a small boy runs out, she hands him a box tied with string, and he lifts the lid and gasps. 24–30 s: the camera rises slowly to show the whole market glowing in the snow. Warm lamplight against blue dusk, falling snow, soft film grain. Snow crunching, a bicycle bell, distant carols, the boy’s laugh.

**Seedance 2.5 FAQ.**
- **What is Seedance 2.5?** Seedance 2.5 is a video model from ByteDance’s Seed team, released on July 31, 2026, as the successor to Seedance 2.0. It generates picture and sound together in clips of 4 to 30 seconds, from a text prompt and reference images, video or audio, and it can cut between linked shots inside one clip. On DreamBench it renders at 480p, 720p or 1080p, with sound on or off.
- **How do I use Seedance 2.5 on DreamBench?** There are 2 ways. On the Video page, pick Seedance 2.5, write your prompt and choose a length from 4 to 30 seconds: your words go to the model as written and come back as one clip. In Film Studio, describe the video you want and add reference images if you have them: DreamBench plans the scenes, has Seedance 2.5 make each one and joins them, and you can redo any scene.
- **How much does Seedance 2.5 cost on DreamBench?** Seedance 2.5 is priced in credits for each second of video: least at 480p, more at 720p and most at 1080p, and the same with sound on or off. A clip on the Video page adds a small base amount; a film adds a base amount for planning it. Credits come with a paid subscription, Seedance 2.5 is on all of them, no ByteDance account or API key is needed, and you see the exact cost before you generate.
- **What is the difference between Seedance 2.5 and Seedance 2.0?** Seedance 2.5 doubles the longest clip from 15 to 30 seconds and takes up to 30 reference images, 10 video clips and 10 audio clips, where 2.0 takes 9, 3 and 3. It follows timings written in whole seconds, where 2.0 follows shot numbers, and ByteDance reworked its skin, eyes, texture and light, so the same prompt looks different on each. Both generate sound and can cut between shots inside a clip.
- **How long can a Seedance 2.5 video be?** One Seedance 2.5 clip runs 4 to 30 seconds. On DreamBench’s Video page you get exactly that: one clip, at the length you pick. In Film Studio a video can run longer, up to 30 seconds on Starter, 2 minutes on Plus and 5 minutes on Ultra and Studio, because DreamBench cuts it into scenes of 3 to 12 seconds and joins them. ByteDance’s own tools can also extend a clip.
- **What resolution does Seedance 2.5 use on DreamBench?** Seedance 2.5 renders at 480p, 720p or 1080p on DreamBench. All 3 are on every subscription, and 1080p is the highest the model renders. 480p costs the least per second, so it suits a first try of an idea or a timing; 720p suits a social clip watched on a phone; choose 1080p for a final cut or a bigger screen. Pick the resolution in the composer before you generate.
- **What aspect ratios does Seedance 2.5 support?** On DreamBench there are 5: square 1:1, widescreen 16:9, vertical 9:16, classic 4:3 and cinematic 21:9. Pick one in the composer, or name it at the start of your prompt and DreamBench sets it for you. ByteDance’s API also lists a taller portrait frame and an adaptive setting that follows the prompt or the input picture; DreamBench doesn’t offer those for Seedance 2.5.
- **Does Seedance 2.5 generate sound and dialogue, and in which languages?** Yes. With sound on, Seedance 2.5 makes dialogue, effects, ambience and music with the picture, timed to what happens on screen; switch it off for picture only, at the same price. BytePlus lists 11 languages for prompts and speech: Chinese, English, Spanish, Indonesian, Malay, Thai, Arabic, Portuguese, Vietnamese, Japanese and Korean. Put each spoken line in quotation marks and name the language it is spoken in.
- **Can Seedance 2.5 make several shots in one clip?** Yes. One Seedance 2.5 clip can cut between linked shots, and ByteDance pitches it at a short story with a set-up, a turn and an ending, the main subject keeping its look across the cuts. Number the shots and give each its seconds, such as “Shot 2 (4–8 s)”. On DreamBench’s Video page your shots reach the model as you wrote them; in a film, a scene’s clip can cut too.
- **How many reference images can I use with Seedance 2.5?** On DreamBench, up to 8 per video, each tagged as a character, creature, location or prop and named in your prompt with @ and its label. DreamBench builds a sheet from each and sends it with every scene it appears in. ByteDance’s API takes up to 30 images, 10 video clips and 10 audio clips in one generation. The Video page takes words only, so references belong in Film Studio.
- **Can I use a photo of a real person with Seedance 2.5?** No. ByteDance doesn’t let Seedance 2.5 take a reference that shows a real person’s face, to protect people’s likeness, and DreamBench suggests another model when one is refused. ByteDance accepts faces from its own models, from preset digital characters and from real people who verified and authorized their likeness. So describe the person instead: if Seedance flags a character DreamBench designed from your words, DreamBench registers it with ByteDance’s asset library and tries again.
- **Can Seedance 2.5 turn an image into a video?** Yes. The model can start a clip from an image as its first frame, or work from reference images of a character, product or place. On DreamBench, add your images in Film Studio and tag them: DreamBench builds a sheet from each and sends it with every scene, so your product, place or character appears where your prompt puts it. The Video page takes words only, and no reference may show a real face.
- **Does Seedance 2.5 support first and last frames?** The model does: ByteDance’s API takes 1 image as the first frame, or 2 as the first and last frames, and the video takes the first image’s shape. DreamBench’s composer has no place to add your own first or last frame. Add the picture as a reference image instead and DreamBench builds the scenes from it, and describe in your prompt how the clip should end.
- **Are Seedance 2.5 videos watermarked?** On DreamBench, videos made on Starter carry a small DreamBench watermark, and Plus, Ultra and Studio remove it. ByteDance’s API can add a mark in the lower-right corner saying a video is AI-generated, an option that stays off unless asked for, and DreamBench doesn’t ask for it. The watermark depends only on your DreamBench subscription, not on the resolution or the length.
- **Can I use Seedance 2.5 videos commercially?** Yes. DreamBench’s Terms of Use allow personal and commercial use of what you make, including ads, product videos and client work. You need the rights to anything you upload, such as product photos or a logo, and generated videos may not be exclusive to you or protected by copyright. Starter videos carry a DreamBench watermark, which Plus and above remove.
- **How do I write a good Seedance 2.5 prompt?** ByteDance suggests a one-line summary first, then the action shot by shot or second by second, then what stays the same throughout. Open with the frame and the length, number the shots with their seconds, give each a size and a camera move, describe the light and the look, and end with the sound, spoken lines in quotation marks. Ask for “no subtitles” or “no music” if you want none.
- **Why doesn’t my Seedance 2.5 video match my prompt?** Usually too much happens in too little time: ByteDance warns that a crowded stretch can come out with extra cuts or missing beats. Give each shot one main action and a few seconds, keep the motion simple, and avoid several people interacting at once, which ByteDance says is not yet fully stable. Lettering can come out misspelt. In a film, redo the scene that missed and keep the rest.
- **How do I keep characters consistent across shots in Seedance 2.5?** Inside one clip, describe the subject once and add a line asking for the same subject, place and light in every shot. Across a film, give DreamBench a reference image of each character, tagged as a character, and name them with @ in your prompt; DreamBench builds a sheet from each and sends it with every scene they appear in. Keep each character’s description the same throughout.
- **Does Seedance 2.5 have a draft mode?** ByteDance’s API has one: it makes a 480p preview, then turns that same preview into a final video. DreamBench doesn’t use it. Instead, 480p is a resolution you can choose for any Seedance 2.5 video, at the lowest price per second, to test an idea; then generate it again at 720p or 1080p. That is a new take, so the idea carries over but the details can differ.
- **How does Seedance 2.5 compare with Kling 3.0 and Veo 3.1 on DreamBench?** They do different jobs. Seedance 2.5 runs to 30 seconds a clip and can cut between shots inside it; Kling 3.0 runs to 15 seconds in one continuous take, with lip-synced dialogue in 5 languages; Veo 3.1 runs to 8 seconds, always with sound, in widescreen or vertical. Seedance 2.5 and Veo 3.1 are on the Video page, and Kling 3.0 runs in Film Studio.
- **Which DreamBench plans include Seedance 2.5?** Available on every plan. Video length depends on the plan: up to 30 seconds on Starter, 2 minutes on Plus and 5 minutes on Ultra and Studio.

**Sources.**
- Seedance 2.5 on ByteDance Seed: https://seed.bytedance.com/en/seedance2_5
- ByteDance’s Seedance 2.5 launch post: https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5
- Seedance 2.5 tutorial (BytePlus ModelArk): https://docs.byteplus.com/en/docs/ModelArk/2607688
- Seedance 2.5 prompt guide (BytePlus ModelArk): https://docs.byteplus.com/en/docs/ModelArk/2607689
- Portrait videos with Seedance (BytePlus ModelArk): https://docs.byteplus.com/en/docs/ModelArk/seedance-portrait-asset-guide

Seedance 2.5 is developed by ByteDance Seed (https://seed.bytedance.com/en/). DreamBench is an independent platform that offers Seedance 2.5 alongside other models and isn't affiliated with or endorsed by ByteDance Seed. Seedance 2.5 and ByteDance Seed are trademarks of their owners, used here only to name the model and its developer. Released 2026-07. Facts last checked 2026-10-31.