SWE-1.6 Benchmark Details
SWE-1.6 currently shows benchmark results led by Vibe Code Bench v1.1 (98 / 112, score 7.22), SWE-bench Multilingual (36 / 40, score 58.30), FrontierCode 1.1 Main (49 / 50, score 9.40).
Benchmark Results
SWE-1.6
Benchmark Results
Repository Engineering
1 evaluationsBenchmark / mode
Score
Rank/total
Code Generation & Editing
2 evaluationsBenchmark / mode
Score
Rank/total