AI BriefingMiMo-V2.6-Distill-Qwen-9B: SWE-bench 61 Score, 9B-class Multimodal Agent
XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B
About the project
This is a lightweight multimodal model distilled from Qwen3.5-9B. It is designed to simultaneously understand text and images, performing complex coding tasks and external tool calls.
It achieved 61.1% on SWE-bench Verified and 44.6% on SWE-bench Pro, demonstrating exceptional coding performance among 9B-class models. It also scored highly on the Terminal-bench 2.1 and Toolathlon benchmarks, proving its practical capabilities as an agent.
Beyond simple chat, it is suitable for agent workflows requiring screenshot analysis or terminal command execution. Licensed under MIT, it can be used without restrictions in commercial projects.
XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B
The original page has no description.
image-text-to-text
This introduction was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report errors, attribution issues, or removal requests via Contact.