Xiaomi is the latest company to release an open-weight AI model \u2013 MiMo-V2.5 claims it is \u201ca major step forward in agentic capability and multimodal understanding.\u201d<\/p>\n
Xiaomi has shared various benchmark results that compare the MiMo-V2.5 to the recently released DeepSeek-V4, Kimi K2.6, Cloud Opus 4.6, Gemini 3.1 Pro, and Xiaomi’s older MiMo-V2-Pro.<\/p>\n
The company claims that MiMo-V2.5 has achieved best-in-class performance on its in-house agentive task benchmark. On the internal MiMo coding bench, the smaller V2.5 model matches the larger V2.5-Pro \u200b\u200bat half the price. Xiaomi says that in other benchmarks testing the model’s image and video understanding, V2.5 is at the level of closed-source models.<\/p>\n
MiMo-V2.5 was evaluated on coding and agentic tasks.<\/strong><\/span><\/p>\n The model was trained on 48 trillion tokens and is natively multimodal with support for text, image, and video data. Xiaomi has published two versions: MiMo-V2.5 with 310B total parameters (15B active) and MiMo-V2.5-Pro \u200b\u200bwith 1.02T total parameters (42B active). The model supports up to 1 million tokens of reference.<\/p>\n You can download the model from here hugging face<\/a> And run it yourself, but you’ll need something like a kitted-out Mac Studio to do it – consumer GPUs don’t have enough VRAM (no, not even the Nvidia RTX 5090).<\/p>\n You can try Xiaomi MiMo-V2.5 AI Studio<\/a> (which is not loaded at the time of writing) or use this via official api<\/a>. Or, as mentioned above, download it and run it locally, if you have the means to do so.<\/p>\n![\"MiMo-V2.5](\"https:\/\/fdn.gsmarena.com\/imgroot\/news\/26\/04\/xiaomi-mimo-v25-model-released\/inline\/-1200\/gsmarena_002.jpg\"\/)
\nMiMo-V2.5 evaluated on image and video understanding<\/strong><\/span><\/p>\n