Quantitative reasoning and autonomous analysis on BixBench-Verified-50 Full set
90AccuracyClaude Code
Evaluation results
| Method | Links | |
|---|---|---|
| Claude CodeCore model=Claude Opus 4.72026.05 | 90 | |
| BIORESEARCHERCore model=GPT-5.4, Reasoning effort=medium2026.05 | 89.33 | |
| CellTypeCore model=Claude Opus 4.72026.05 | 89.33 |