ByteDance Doubao Large Model 2.0 scheduled for official release on February 14

ByteDance’s Volcano Engine has tentatively scheduled the release of a series of major upgrades for the Doubao large model on February 14, 2026. This model release includes Doubao Large Model 2.0, Audio and Video Creation Model Seedance 2.0, and Image Creation Model Seedream 5.0 Preview. With this upgrade, Doubao Large Model 2.0 will be officially launched, significantly enhancing its foundational model capabilities and enterprise-level Agent capabilities. Key points of the Doubao Video Generation Model Seedance upgrade include high usability for complex interactions and motion generation, rated as industry best; comprehensive multimodal capabilities supporting audio, video, and image inputs; strong controllability with excellent instruction adherence; deep adaptation to film, advertising, and marketing scenarios, with output quality aligned to industrial delivery standards. Additionally, the key points of the Doubao Image Creation Model Seedream upgrade include the introduction of real-time retrieval enhancement capabilities to access the latest knowledge and information, enabling precise responses to time-sensitive creative needs; enhanced world knowledge and multilingual capabilities, with the model embedded with rich knowledge from science, technology, and humanities fields; and overall improvements in understanding and generation performance, capable of interpreting user intent through brief, vague text and image inputs, with progress in subject consistency and text-image alignment. (Science and Technology Innovation Board Daily)