模型详情
dreamina-seedance-2-5-260628
Seedance 2.5 is a video generation model from ByteDance. It is suited for long-form storytelling, multimodal reference-based generation, video editing, and video extension. It supports first-frame and first-and-last-frame control, up to 50 image, video, and audio reference assets, optional generated audio, and multilingual audiovisual generation.
模型规格
- 上下文长度
- –
- 最大输出
- –
- 输入/输出模态
- –
- 发布日期
- 2026-08
02
API 端点
单一 MixRoute 网关，OpenAI 兼容
-
OpenAI-compatible
/v1/chat/completionsPOST
03
阶梯价格
单位：/unit
| 档位 | 价格 /unit |
|---|---|
| 1080p_video res == "1080p" && has_video | $7.0000 |
| lowres_video (!(res == "1080p" && has_video)) && (has_video) | $6.4000 |
| 1080p (!(res == "1080p" && has_video)) && (!(has_video)) && (res == "1080p") | $11.7000 |
| lowres (!(res == "1080p" && has_video)) && (!(has_video)) && (!(res == "1080p")) | $10.7000 |
05
选型摘要
快速判断是否适合你的工作负载
What is this model good for?
Seedance 2.5 is a video generation model from ByteDance
Seedance 2.5 is a video generation model from ByteDance. It is suited for long-form storytelling, multimodal reference-based generation, video editing, and video extension.
What should you check before using it?
Check duration, resolution, and aspect ratio limits
Model type: video. Verify supported features and quotas with the provider documentation. Released 2026-07-31.
Why use it through MixRoute?
Use the compatible endpoint with stable model ID
Use the model ID dreamina-seedance-2-5-260628 through MixRoute, then run the same integration tests used for the current client.
单一端点，可验证的决策
使用相同的请求格式测试此模型和备用路由
从真实工作负载开始，再由质量、总成本和故障条件决定是否接入生产流量。