ByteDance Launches Seedance 2.5: Supports 50 References and 30s Video Generation
FellMentKE · x · 2026-08-02
ByteDance's AI product Dreamina has globally launched the video generation model Seedance 2.5. The model aims to eliminate the traditional trade-offs creators face between freedom and control, or reference volume and workflow manageability.
Key features include:
- Support for up to 50 multimodal references
- Single-generation video up to 30 seconds
- The lowest-cost access to the Seedance family
Early access is currently live on Dreamina, with US availability expected in about a week.
Related event: ByteDance Seedance 2.5: 50 refs, 30s video, CapCut integration(33 posts)→
More from Multimodal
- Split video generation from assembly: render captions and graphics locally to cut API costs — ziggidyy_01 · 2026-09-16
- User hails astra's image handling as the most fun they've had yet — TAbrodi · 2026-09-16
- Video model ranking sparks debate: Seedance 2.5 takes #1, Wan 3.0 #2, Seedance 2.0 #3 — Aiden_Tech_Ai · 2026-09-16
- DiffusionLab Curates Vintage SD 1.5 Art From the Tesla K80 Era — cjimti · 2026-09-16
- Dev builds open-source real-time insurance claim voice agent with Gemini, handles mid-call language switch — Saboo_Shubham_ · 2026-09-16
- NVIDIA's FlashDreams Gives Autoregressive Video and World Models an Inference Runtime — techNmak · 2026-09-16