Scientific Curation Benchmark
81GRAS ScoreBeaver
Evaluation Results
| Method | Links | |
|---|---|---|
| BeaverModel=gpt-5.42026.06 | 81 | |
| CodexModel=gpt-5.42026.06 | 58.6 | |
| Claude CodeModel=opus-4.72026.06 | 58.5 | |
| Claude CodeModel=opus-4.62026.06 | 58 | |
| Copilot CLIModel=gpt-5.42026.06 | 57.4 | |
| Meta-HarnessModel=gpt-5.42026.06 | 48 | |
| Claude CodeModel=haiku-4.52026.06 | 44 |