3 ms·

Benchmark geometric mean - GPT-5.5: 62.7% - Opus 4.8: 62.2% - Kimi K2.7 Code: 56.3% - Kimi K2.6: 48.2%

• lostmsu 4mo ago
  Would be nice to have 5.2 and 4.6 for comparison.