Dreamina Seedance 2.5
#22 v Text na videoByteplus · v2.5 · od 31. Juli 2026 (weltweiter Launch auf Dreamina) · 2× · naposledy 18. 8. 2026
33
Momentum
Dreamina Seedance 2.5 is an AI video model from ByteDance/BytePlus, available via the Dreamina platform and the BytePlus API (ModelArk). It generates continuous video scenes up to 30 seconds long (up to 180 seconds in beta long-video mode) from text, image, video and audio inputs, supporting up to 50 multimodal reference assets for precise control over characters, style and motion. The model launched globally on Dreamina on July 31, 2026; API pricing varies by resolution (480p/720p) and whether video input is used.
Vývoj momenta
20.05.18.08.
Vlastnosti
| Fine-Tuning | No fine-tuning; instead control via up to 50 multimodal reference assets (image, video, audio) per prompt |
| Max Resolution | Per official BytePlus page: max 720P; Dreamina consumer page additionally advertises 4K (not corroborated by API rate card) |
| Max Video Length | Up to 30 seconds in standard mode, up to 180 seconds in beta long-video mode |
| Platform | Dreamina (dreamina.capcut.com), BytePlus ModelArk API, Volcano Engine Ark (China) |
| Price | API: $10.70 / 1M tokens (no video input), $6.40 / 1M tokens (with video input); example: $0.514 for 5s at 480p |
| Release Date | Global launch on Dreamina on July 31, 2026 |