MiMo-V2-Flash
A Mixture-of-Experts model positioned around fast reasoning, coding, long context, and agentic workflows.
309B total parameters15B active parameters256k context
Source: MiMo-V2-Flash README
Models overview
Xiaomi MiMo models
The public MiMo line spans earlier reasoning-first 7B releases and the newer MiMo-V2-Flash family. This directory keeps the main public variants readable in one place.
MiMo-V2-Flash-Base
The base MiMo-V2-Flash variant used as the public foundation for benchmark comparisons and downstream evaluation.
309B total parameters15B active parameters256k context
Source: MiMo-V2-Flash model table
MiMo-7B-RL
The reasoning-focused 7B public line from the earlier MiMo release, positioned around math and code performance.
7B class modelRL-tunedReasoning and code focus
Source: MiMo README
MiMo-7B-Base
The earlier base model from the MiMo line, useful for understanding the public evolution from MiMo-7B to MiMo-V2-Flash.
7B class modelPre-training heavyReasoning-oriented
Source: MiMo README