ChartMuseum test split, 1,162 expert-annotated questions; average over 5 runs.
条件
Adaptive thinking/max effort, no tools. Internal implementation follows public prompts but uses Claude Sonnet 4.6 grader instead of GPT-4.1-mini; retain grader/config as a distinct comparison.
ChartMuseum test split, 1,162 expert-annotated questions; average over 5 runs.
条件
Adaptive thinking/max effort with Python tools. Internal implementation uses Claude Sonnet 4.6 grader instead of GPT-4.1-mini; retain grader/config as a distinct comparison.
Claude Opus 4.8 · ChartMuseum · 1,162 test-split questions · Scores and configurations
Claude Opus 4.8 record for ChartMuseum · 1,162 test-split questions. 2 score records for this release, including 2 exact values (may include historical records). Includes 2 source locations.
ChartMuseum test split, 1,162 expert-annotated questions; average over 5 runs.
Conditions
Adaptive thinking/max effort, no tools. Internal implementation follows public prompts but uses Claude Sonnet 4.6 grader instead of GPT-4.1-mini; retain grader/config as a distinct comparison.
ChartMuseum test split, 1,162 expert-annotated questions; average over 5 runs.
Conditions
Adaptive thinking/max effort with Python tools. Internal implementation uses Claude Sonnet 4.6 grader instead of GPT-4.1-mini; retain grader/config as a distinct comparison.