ByteDance launches Seedance 2.5 Flash video model at about 0.99 yuan per second for 720p, 34% below Seedance 2.5
ByteDance's Volcano Engine released Seedance 2.5 Flash on November 12, pricing 720p text-to-video and image-to-video at about 0.99 yuan per second, roughly 34% below Seedance 2.5. The company claims it generates video twice as fast as Seedance 2.5. API access on Volcano Ark opens November 15.
ByteDance's cloud unit Volcano Engine released Seedance 2.5 Flash on November 12, a lighter version of its Seedance 2.5 video generation model, priced at about 0.99 yuan per second of 720p video. Volcano Engine puts that roughly 34% below Seedance 2.5 for text-to-video and image-to-video. The company claims the model generates video twice as fast as Seedance 2.5.
The details come from reporting by Pandaily, which cites IT Home and the official BytePlus Seedance 2.5 Flash page. ByteDance writes the name as "Seedance 2.5 flash" in Chinese materials, while its BytePlus international site styles it Seedance 2.5 Flash.
Specifications
- Price: about 0.99 yuan per second of generated video at 720p, for text-to-video and image-to-video
- Clip length: up to 30 seconds
- Output resolution: 480p or 720p
- Reference inputs per request: up to 10 images, five video clips and five audio clips
- Editing: supports video editing, so users can adjust content and produce different creative versions
Quality claims
Volcano Engine says the model performs better overall than Seedance 2.0, the previous flagship. It also says the model maintains a high standard for instruction following, physical motion and audio, and that its output is the closest of any model to the flagship Seedance 2.5. These are company claims. ByteDance did not publish benchmark results or side-by-side comparisons with rival models.
Two-tier family
The release splits the Seedance 2.5 family into two tiers. The flagship Seedance 2.5 remains the choice for the highest audiovisual quality and serves 1080p and 4K production. It is paired with a Draft mode that lets users pick a preview clip before generating the final high-resolution version, which cuts the cost of trial and error.
Seedance 2.5 Flash is pitched at high-frequency creation and batch content production, where speed and cost matter more than the last increment of quality.
Availability
Seedance 2.5 Flash is rolling out from November 12 on ByteDance's Jimeng creative platform and on third-party tools including LibTV, Updream, Creao AI and TapNow. API access is scheduled to open on Volcano Ark, Volcano Engine's model platform, on November 15.
Seedance 2.5, which ByteDance detailed this summer, introduced 30-second single generations and multimodal references. Volcano Engine did not say whether the flash model's lower price will extend to 1080p output, or whether flagship pricing will change.
What this means
The release gives Seedance 2.5 users a split between cost and fidelity. Flash caps output at 720p, while the flagship covers 1080p and 4K. Teams producing clips in volume can use the cheaper model, and those needing the highest resolution stay on Seedance 2.5, previewing with Draft mode before committing to a final render.
The 30-second clip length and the multimodal reference limits (10 images, five video clips, five audio clips) match the capabilities that Seedance 2.5 introduced, so Flash carries the main workflow features of the flagship at the lower tier. The speed and quality-closeness claims are unverified, since no benchmarks or comparisons accompany them. Developers should test output on their own workloads once Volcano Ark API access opens on November 15.
Related Articles
Baidu launches ERNIE Lightning reasoning model and ERNIE-Wizshot-ME video model on Qianfan
Baidu released ERNIE Lightning, a deep-thinking model built first for Baidu Search, and ERNIE-Wizshot-ME, an expressive video model, at Baidu World 2026 on November 12. Both are live on Baidu AI Cloud's Qianfan platform with APIs open to enterprise customers and partners. Baidu has not published benchmarks, parameter counts or context lengths for either.
Geodesic Intelligence releases NovaAtom-Lite, a biomolecule model it claims tops state-of-the-art systems
Geodesic Intelligence has released NovaAtom-Lite, an AI model that predicts the structures and interactions of biomolecules. Founder Quanquan Gu said the model topped a third-party benchmark in several drug discovery tasks, according to the South China Morning Post.
Tether releases QVAC BotPsy-0.8B, an open-weight 0.8B model for tool calls on personal devices
Tether AI Research announced QVAC BotPsy-0.8B on November 11, 2026. The open-weight, 0.8-billion-parameter model is built on Alibaba's Qwen3.5-0.8B and targets function calling and multi-step agentic workflows on personal devices.
Claude Fable 5.2 scores 65 on Artificial Analysis index but costs about 2x Opus 5.5 per task
Anthropic announced Claude Fable 5.2, its top model, on November 9 (US time). ITmedia tested it against Opus 5.5 and Sonnet 5.5 on two office tasks and found the best-structured output, at the highest cost.
Comments
Loading...