ByteDance Douyin Big Model 2.0 is set to be officially released on February 14th.

ByteDance's Volcano Engine has tentatively scheduled a series of important upgrades for the release of the Dobag Big Model on February 14, 2026. This model release involves the Dobag Big Model 2.0, the audiovisual creation model Seedance 2.0, and the image creation model Seedream 5.0 Preview. In this upgrade, the Dobag Big Model 2.0 will be officially released, with significantly improved basic model capabilities and enterprise-level Agent capabilities. Key upgrades to the Dobag video generation model Seedance include high complexity in interactions and operation, with the highest level of availability in the industry; comprehensive multi-modal capabilities, supporting audio, visual, and image inputs in all modes; strong controllability, with instructions following good performance; deep adaptation to film, advertising, and marketing scenes, aligning output quality with industrial delivery standards. Additionally, key upgrades to the Dobag image creation model Seedream include the first introduction of real-time retrieval enhancement capabilities, allowing for the retrieval of the latest knowledge and information to accurately respond to time-sensitive creative needs; enhanced world knowledge and multilingual capabilities, with the model containing rich world knowledge in technology and humanities fields; overall improvement in understanding and generation performance, able to understand user intent through short, fuzzy text and image inputs, with improved consistency of subjects and alignment of text and images.