Home Model series Claude Opus Claude Opus 4.6 Claude Opus 4.6 · MMMU-Pro · Standard and Vision formats · Scores and configurations Claude Opus 4.6 · MMMU-Pro · Standard and Vision formats · Scores and configurations Claude Opus 4.6 record for MMMU-Pro · Standard and Vision formats. 2 score records for this release, including 2 exact values (may include historical records). Includes 2 source locations.
Scores and evaluation conditions Claude Opus 4.6 · MMMU-Pro scores and configurations Metric and score Configuration and evaluation conditions Sources Accuracy 73.9% Reported 73.9% · Unit % Published Unconfirmed
Max · 无工具 Settings Adaptive thinking、Max effort、无工具;sampling 使用默认 temperature 与 top_p。 Unknowns none recorded Tools Tools forbidden Reasoning budget max Framework not recorded Protocol MMMU-Pro;without tools;Standard 10-option 与 Vision 格式。 Conditions 两种格式分别5-run average;adaptive thinking、max effort;无图像裁切工具。 View this score in rankings Accuracy 77.3% Reported 77.3% · Unit % Published Unconfirmed
Max · 图像裁切工具 Settings Adaptive thinking、Max effort;允许简单 image-cropping tool;每项评测条件见成绩记录。 Unknowns none recorded Tools Tools allowed Reasoning budget max Framework not recorded Protocol MMMU-Pro;使用 simple image-cropping tool;Standard 10-option 与 Vision 格式。 Conditions 两种格式分别5-run average;adaptive thinking、max effort。此前 configId 错误复用了 HLE web/code 工具配置,已修正为独立图像裁切配置。 View this score in rankings
Benchmark versions and metrics Aliases MMMU-Pro
Accuracy · % · Higher is better