Seedance
Seedance is ByteDance's family of AI video models; Seedance 2.0 and 2.5 rank 6th and 7th on Arena's text-to-video leaderboard as of Nov 6, 2026, and 2.5 makes up to 30 seconds with sound in one pass. You can use it in ByteDance's Dola and Dreamina, through the BytePlus API, or inside Runway, Luma, Pika and HeyGen.
Seedance is the AI video model family from ByteDance's Seed research team, the company behind TikTok and CapCut, and as of Nov 6, 2026 its newest model, Seedance 2.5, generates up to 30 seconds of video with sound in one generation and can extend a clip twice (ByteDance Seed). Seedance 2.0 and 2.5 rank 6th (1481) and 7th (1480) of 50 on Arena's text-to-video leaderboard, behind Google's Gemini Omni, Black Forest Labs' Flux 3 Video and Grok Imagine Video 1.5 (Arena, read Nov 6, 2026). ByteDance offers them in its Dola assistant, in Dreamina and through the BytePlus ModelArk API, and other apps resell them: Runway charges 160 credits per 4 seconds of Seedance 2.0 Pro at 1080p.
Pick Seedance for long takes with sound, steering by reference videos, and a model that ranks near the top. Google's Gemini Omni in Google Flow still scores higher (1st, 1541) and has a free daily allowance; Kling 4.0 matches the 30-second length; and if you need open weights, look at MiniMax's H3 or Wan, since Seedance has none. Best AI Video Generator compares them.
Quick facts
| Get it | Dola, Dreamina, BytePlus ModelArk API; inside Runway, Luma, Pika and HeyGen |
| Maker | ByteDance Seed |
| Latest model | Seedance 2.5 (in Luma by Aug 6, 2026) |
| Other models | Seedance 2.0 (Pro, Fast, Mini), 1.5 Pro, 1.0 Pro, 1.0 Lite |
| Clip length | Up to 30 seconds, extendable twice (2.5) (ByteDance) |
| Audio | Generated with the video (2.0 and 2.5) |
| Ranking | 2.0 6th (1481), 2.5 7th (1480) on Arena text-to-video (read Nov 6, 2026) |
| Open weights | No |
Models
Seedance 2.5 is described by ByteDance as a joint audio and video model with smoother motion, more realistic images and more dependable editing; it reads reference videos for framing and camera language, not just motion, and adds production controls such as untextured "white model" control, green-screen editing, camera movement and performance blocking (ByteDance Seed). It reached Luma on Aug 6, 2026 and Runway's MCP on Aug 11, 2026.
Seedance 2.0 takes text, images, audio and video as input and lets you direct performance, lighting, shadow and camera movement, with all three of images, audio and video usable as references (ByteDance Seed). Runway added it on Apr 7, 2026, the same day HeyGen announced a Seedance 2.0 integration. Mini and Fast tiers followed in Runway on Jul 21, 2026.
Seedance 1.0, ByteDance's first, generates multi-shot 1080p video from text or images; its published benchmark is dated Jun 9, 2025 (ByteDance Seed). Seedance 1.5 Pro and 1.0 Pro rank 24th and 37th on Arena today.
Where to use Seedance and what it costs
ByteDance's own apps. ByteDance Seed's model pages link to its Dola assistant and to the BytePlus Playground for trying the models, and Arena lists them as "Dreamina Seedance," Dreamina being ByteDance's creative app (Arena).
API. BytePlus ModelArk sells Seedance 2.5, 2.0, 2.0 Fast and Mini, 1.5 Pro and 1.0 Pro to developers; its pricing page was last updated Oct 23, 2026 but did not show its rates in a readable form on Nov 6, 2026 (BytePlus).
Other apps. Runway sells Seedance 2.0 Pro at 1080p for 160 credits per 4 seconds, or 40 credits a second, more than three times its own Gen-4.5's 12 (Runway). Luma offers Seedance 2.0, 2.5 and 2.0 Mini, Pika offers 2.5, and HeyGen pairs Seedance 2.0 scenes with its avatars.
What a creator saw. In Dan Kieft's Oct 16, 2026 guide, made in Higgsfield (which sponsored it), Seedance 2.5 at 1080p gave the most realistic of his four test clips without references, against MiniMax H3 Max and Flux 3 Video, and he called it "quite expensive"; its cheaper 480p draft mode let him regenerate until he liked a clip, then upscale only that one to 1080p.
Seedance vs the alternatives
Google Flow and Veo: Gemini Omni 1.2 Flash leads Arena at 1541 against Seedance 2.0's 1481, and Flow gives 50 free credits a day. Seedance 2.0 and 2.5 both rank well above Google's older Veo 3.1 (14th).
Kling: Kling 4.0 also makes 30-second takes, with up to 10 keyframes, but has no Arena ranking yet.
Hailuo: MiniMax H3 (11th) is shorter at 15 seconds but has open weights.
Wan: Alibaba's Wan 3.0 is just behind at 8th (1477); older Wan versions can run on your own computer.
Sora: OpenAI's Sora 2 Pro, now discontinued, ranks 13th.
Frequently asked questions
What is the newest Seedance model? Seedance 2.5 is ByteDance's newest video model as of Nov 6, 2026; it makes up to 30 seconds of video with sound in one generation and can extend a clip twice.
Where can I use Seedance? Seedance is available in ByteDance's Dola assistant and Dreamina, through the BytePlus ModelArk API, and inside Runway, Luma, Pika and HeyGen.
Is Seedance better than Veo? On Arena's text-to-video leaderboard read Nov 6, 2026, Seedance 2.0 (6th) and 2.5 (7th) rank above Veo 3.1 (14th), but below Google's newer Gemini Omni 1.2 Flash (1st).
Does Seedance make sound? Yes. Seedance 2.0 and 2.5 generate audio and video together in one pass.
Is Seedance open source? No. ByteDance offers Seedance through its apps and API only; for open-weight video models, see Wan or MiniMax H3.
How much does Seedance cost? In Runway, Seedance 2.0 Pro at 1080p costs 160 credits per 4 seconds as of Nov 6, 2026; ByteDance's own BytePlus rates were not readable that day.
Sources
- Seedance 2.5, Seedance 2.0 and Seedance 1.0, ByteDance Seed, read Nov 6, 2026
- ModelArk pricing, BytePlus, read Nov 6, 2026
- Text-to-video leaderboard, Arena, read Nov 6, 2026
- Pricing and Changelog, Runway, read Nov 6, 2026
- News, Luma AI, read Nov 6, 2026
- Blog, HeyGen, read Nov 6, 2026
- How I Make Realistic AI Videos with Seedance 2.5 - Full Guide, Dan Kieft, Oct 16, 2026 (video)
Moments
How I Make Realistic AI Videos with Seedance 2.5 - Full Guide
Dan Kieft
See also
-
RunwayRunway is an AI video company whose app makes and edits video with its own Gen-4.5 and Aleph 2.0 models plus outside models such as Kling 4.0 and Seedance. It has a free trial with 125 credits, and paid plans cost $15, $35 or $95 a month as of Nov 6, 2026. -
Google FlowGoogle Flow is Google's AI filmmaking app at flow.google.com, built on the Gemini Omni Flash and Veo video models and Nano Banana for images. Everyone 18 or older in a supported country gets 50 free credits a day; Google AI Pro adds 1,000 credits a month and Ultra 10,000 or 25,000, as of Nov 6, 2026. -
KlingKling is the AI video generator from Kuaishou's Kling AI unit; its newest model, Kling 4.0, launched Oct 15, 2026 and makes clips of up to 30 seconds in one pass. The Kling 3.0 API costs from $0.084 a second, and Kling also runs inside Runway, Luma and Pika, as of Nov 6, 2026.
-
HailuoHailuo is MiniMax's AI video app at hailuoai.video; its current model, MiniMax H3 (Jul 31, 2026), makes 4 to 15 second videos with stereo sound at up to 2K and has open weights. H3 ranks 11th on Arena's text-to-video leaderboard as of Nov 6, 2026.
-
WanWan is Alibaba's family of AI video models; Wan 2.1 and Wan 2.2 are free open models under the Apache 2.0 license that run on a home graphics card, while the newer Wan 2.5 to 2.7 and Wan 3.0 are offered only through Alibaba's apps and API as of Nov 6, 2026. -
Best AI Video GeneratorThe best AI video generator in Nov 2026 on LMArena's votes is Google's Gemini Omni Flash, and on Artificial Analysis's it is Alibaba's Wan 3.0; ByteDance's Seedance and MiniMax's Hailuo models rank near the top of both. OpenAI closed the Sora app in 2026, and Sora 2 and Veo 3.1 now rank 13th and 14th on LMArena.