xiaomi
* This model's size is not published. The figure is the source's own estimate, inside a reported interval of 220B–2.3T · confidence: reported .
Knowledge-probe estimate from the paper's leaderboard (revision v2): the figure is the point estimate, the interval is its 90%% prediction interval.
MiMo-V2.5 is a native omnimodal model by Xiaomi. It delivers Pro-level agentic performance at roughly half the inference cost, while surpassing MiMo-V2-Omni in multimodal perception across image and video understanding...
The OpenRouter API truncates descriptions — read the full text at the source
Every id and name this model has been seen under — for matching it against another catalogue.