Benchmark · KnowledgeSettled

SWE-bench Pro

Updated 2026-04-07
Models tested
3
Top score
77.8
Claude Mythos Preview
Median
77.8
min 77.8
Top-5 spread
σ 0.0
Settled
SWE-BENCH PRO \u00B7 TOP 30255075100#1Claude Mythos Preview77.8#2Claude Opus 4.7VERIFIED64.3#3Claude Opus 4.6VERIFIED53.4benchgecko.ai/benchmark/swe-bench-pro

Same category · related evaluations