Bongard Problem Verification on Bongard Problems BPs #2–#100 (test)
86.5AccuracyClaude 3.7 Sonnet
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude 3.7 Sonnetverifier=natural language + programs2026.02 | 86.5 | |
| Claude 3.7 Sonnetverifier=natural language2026.02 | 83.5 | |
| Claude 3.7 Sonnetverifier=programs2026.02 | 82.1 | |
| GPT-4overifier=natural language + programs2026.02 | 79 | |
| GPT-4overifier=natural language2026.02 | 77.5 | |
| GPT-4overifier=programs2026.02 | 72.7 |