XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B

Xiaomi MiMo releases a 9B agentic model fine-tuned from Qwen3.5-9B on 77.4B tokens. The checkpoint targets coding, cybersecurity, and general agent tasks, with documented SGLang deployment steps. Reported benchmarks show gains over the base model, though internal evaluation sets limit independent verification of specific performance claims.

README

XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B View on Hugging Face

Loading the README from Hugging Face…