Xiaomi Open-Sources MiMo-V2.5 AI Models Under MIT License
Xiaomi has open-sourced its MiMo-V2.5 model family under the MIT License, allowing commercial deployment, continued training and fine-tuning without additional authorization. The release includes MiMo-V2.5-Pro, which Xiaomi positioned for complex agent and coding tasks, and MiMo-V2.5, which it described as a native omni-modal model with agent capabilities. Both models support a 1 million-token context window and are available in base and instruct versions.
MiMo-V2.5-Pro has 1.02 trillion total parameters with 42 billion active parameters, while MiMo-V2.5 has 310 billion total parameters with 15 billion active parameters. Xiaomi said the Pro model ranks first among open-source models on GDPVal-AA and ClawEval, the broader MiMo-V2.5 series has support in SGLang and vLLM, and MiMo-V2.5 was trained on about 48 trillion tokens using FP8 mixed precision.
From the sources (6 posts)
@teortaxestex…It wasn't an intern's joke MiMo 2.5 (not Pro): > Trained on a total of ~48T tokens using FP8 mixed precision. The context window supports up to 1M tokens. We've got another 1M class, and the largest disclosed pretrain. Congrats Xiaomi.
@lmsysorgRT @XiaomiMiMo: SGLang and vLLM support for the MiMo-V2.5 series is here. 🙌 Huge thanks to SGLang project from @lmsysorg and @vllm_project…
@teortaxestexRT @teortaxesTex: This is a big F U to DeepSeek and frankly other labs that work with fancier architectures. For the next generation Xiaomi…
@theahmadosmanNew Opensource SoTA contender enters the arena Xiaomi MiMo-V2.5 Pro - 1.02T Total Params / 42B Active Params - Base and Instruct versions Xiaomi MiMo-V2.5 - 310B Total Params / 15B Active Params - Base and Instruct versions MIT License
@jeremyphowardRT @XiaomiMiMo: Xiaomi MiMo-V2.5 is now officially open-sourced! MIT License, supporting commercial deployment, continued training, and fin…
@teortaxestexRT @_LuoFuli: Just dropped two open-source models: MiMo-V2.5-Pro (Code Agent, 1T total) and MiMo-V2.5 (Multimodal Agent, 310B total). Oh a…