B.AI Adds Another Major Free Model: Xiaomi MiMo-V2.5 Fully Multimodal Flagship Now Officially Available

The B.AI platform announced that Xiaomi's open-source flagship multimodal model, MiMo-V2.5, is now officially available for fully free API access. As a 310B-parameter sparse MoE native multimodal large model, MiMo-V2.5 supports comprehensive multimodal input across text, images, video, and audio. Featuring a 1M ultra-long context window and ultra-fast response speeds, it is deeply optimized for high-intensity real-world scenarios such as multimodal agents, audio-visual intelligent analysis, and general-purpose programming. Starting today, developers can unlock the full power of Xiaomi's cutting-edge multimodal model at zero cost through the official API channel on the B.AI platform. With no barriers to entry, you can seamlessly integrate top-tier multimodal capabilities directly into your own applications and workflows. Visit chat.b.ai/chat right now to experience the next generation of multimodal intelligence powered by Xiaomi MiMo-V2.5 completely free of charge.