Seed 2.0 Mini

Seed 2.0 Mini is ByteDance's lightweight multimodal language model, part of the Seed family of models. As the compact tier offering in ByteDance's lineup, it provides multimodal capabilities at a reduced computational cost compared to larger models in the family. The model supports text, image, and video inputs with a substantial 262,144 token context window, allowing it to process lengthy documents or multiple media files in a single request. This multimodal capability enables it to analyze visual content, understand video sequences, and generate text responses based on combined text and visual inputs. However, it does not include tool calling functionality. Seed 2.0 Mini is designed for applications requiring multimodal understanding at scale, particularly where cost efficiency is important. Its video processing capabilities distinguish it from text-only lightweight models, making it suitable for content analysis, media understanding, and applications that need to process visual information alongside text.