all the models — AI benchmark observatory
← Benchmarks

MiMo Coding Bench

MiMo Coding Bench evaluates coding-agent capabilities on software engineering tasks reported with the MiMo model family.

id mimo-coding-bench · max 1 · 4 models reported

#ModelScore
1MiMo-V2.5-Pro
Xiaomi · open
0.74
2MiMo-V2.5
Xiaomi · open
0.72
MiMo Coding Bench Leaderboard · all the models