all the models — AI benchmark observatory
← Benchmarks

FrontierSWE (Impl.)

FrontierSWE (Impl.) evaluates software engineering implementation ability and reports model ranking on implementation tasks. Lower rank is better.

id frontier-swe-impl · max 1 · 1 models reported

#ModelScore
1MiMo-V2.5-Pro
Xiaomi · open
3.40
FrontierSWE (Impl.) Leaderboard · all the models