The video generation model "Seedance 2.5" is scheduled to start its full API release on July 16th. This indicates that the model is moving from the gray zone test phase to formal commercialization, further lowering the barrier for high-quality video generation.

Previously, this model was officially released on June 23rd, and it had announced a release in early July, but the specific date was not disclosed. This full release marks the official confirmation of the implementation process. Currently, the video creation platform "Ji Meng (Jimen)" is conducting a pre-release on its official website, and it has announced special events such as member discounts and points campaigns after the official release.

The core of this update is that Seedance 2.5 has achieved technological breakthroughs in three aspects: the length of single-section generation, multiple material references, and video editing. This model can now directly output original videos up to 30 seconds long, significantly expanding the continuous expression space of AI videos. In terms of multimodal fusion, the capacity of reference materials has been increased to a maximum of 50 multimodal materials, enhancing contextual understanding in complex scenes. Additionally, this model introduces more controllable video generation and editing functions, allowing creators to accurately adjust the main subject of images and camera work.