Two Xiaomi MiMo models have officially arrived on the B.AI API, bringing multimodal intelligence and advanced reasoning capabilities to developers. 🔥
First is MiMo-V2.5, a 310B sparse MoE model built with native multimodal capabilities.
It supports:
🔸 Image inputs
🔸 Text inputs
🔸 Audio inputs
🔸 Video inputs
🔸 A 1M context window
Designed for multimodal Agents and general coding, MiMo-V2.5 offers a broad range of capabilities through the API.
Then there is MiMo-V2.5-Pro.
A 1.02T sparse MoE flagship built for complex reasoning and long-horizon software engineering, with a reported score of 78.9 on SWE-Bench Verified.
Both models are now available through the B.AI API.
Official pricing applies, while Custom Provider users receive 60% OFF. 🚀
Get started: https://chat.b.ai/chat
#TRONEcoStar @Justin Sun孙宇晨
First is MiMo-V2.5, a 310B sparse MoE model built with native multimodal capabilities.
It supports:
🔸 Image inputs
🔸 Text inputs
🔸 Audio inputs
🔸 Video inputs
🔸 A 1M context window
Designed for multimodal Agents and general coding, MiMo-V2.5 offers a broad range of capabilities through the API.
Then there is MiMo-V2.5-Pro.
A 1.02T sparse MoE flagship built for complex reasoning and long-horizon software engineering, with a reported score of 78.9 on SWE-Bench Verified.
Both models are now available through the B.AI API.
Official pricing applies, while Custom Provider users receive 60% OFF. 🚀
Get started: https://chat.b.ai/chat
#TRONEcoStar @Justin Sun孙宇晨
