Xiaomi (小米) Launches MiMo-V2.6 Model Series on B.AI
Xiaomi (小米) has released its MiMo-V2.6 series of open-source native multimodal reasoning models on the B.AI platform, available via both API and Web Chat. The series features a 1M token context window and comprehensive understanding of text, image, video, and audio. The MiMo-V2.6-Pro model utilizes a sparse MoE architecture with 1.02 trillion total parameters, while the MiMo-V2.6-Flash model is optimized for high throughput and rapid response.
Summaries are written by AI from the original article. Not investment advice.