Scientific QA on SCIMDR Eval
49.9Accuracy
GPT-5.2
Updated 4mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.03 | 49.9 | |
2026.03 | 49.1 | |
2026.03 | 47.2 | |
2026.03 | 34.2 | |
2026.03 | 19.8 | |
2026.03 | 16.8 | |
2026.03 | 15.5 | |
2026.03 | 5.6 |
GPT-5.2
| Method | Links | |
|---|---|---|
2026.03 | 49.9 | |
2026.03 | 49.1 | |
2026.03 | 47.2 | |
2026.03 | 34.2 | |
2026.03 | 19.8 | |
2026.03 | 16.8 | |
2026.03 | 15.5 | |
2026.03 | 5.6 |