Claude Mythos Preview Leads SWE-bench Verified at 93.9%, Opening 6.3-Point Gap Over Claude Opus 4.7

BenchLM·medium signal

As of May 22, 2026, Claude Mythos Preview tops the SWE-bench Verified leaderboard at 93.9%, followed by GPT-5.5 at 88.7% and Claude Opus 4.7 Adaptive at 87.6%. The 6.3-point gap between Mythos Preview and Opus 4.7 suggests a significant capability jump in agentic coding. With 47 models now evaluated, the leaderboard shows the frontier has moved from ~80% to ~94% in under a year — approaching human-level on real GitHub issue resolution.