BioSimBench 1.0: New drugs and cell contexts
Science
Perturbation-specific F1 on 60 held-out cases predicting gene expression changes after a drug, with the drug, the cell line or both held out.
Benchmark source: Celltype
Top models (higher is better)
| Model | Score | Source |
|---|---|---|
| Gemini 4 Argon | 31.8 | BioSimBench 1.0: Gemini 4 Argon · CellType |
| Claude Sonnet 5.5 | 28.6 | BioSimBench 1.0: Claude Sonnet 5.5 · CellType |
| GPT-6 Astra | 25.7 | BioSimBench 1.0: GPT-6 Astra · CellType |
| Claude Opus 5.5 | 24.7 | BioSimBench 1.0: Claude Opus 5.5 · CellType |
| GPT-6.1 Sol | 23.4 | BioSimBench 1.0: GPT-6.1 Sol · CellType |
| Gemini 3.8 Flash | 23.0 | BioSimBench 1.0: Gemini 3.8 Flash · CellType |
| Gemini 3.1 Pro Preview | 22.2 | BioSimBench 1.0: Gemini 3.1 Pro · CellType |
| GPT-6 Luna | 18.7 | BioSimBench 1.0: GPT-6 Luna · CellType |
| Grok 4.7 | 18.7 | BioSimBench 1.0: Grok 4.7 · CellType |
| GPT-6 Sol | 18.5 | BioSimBench 1.0: GPT-6 Sol · CellType |
| Claude Haiku 5.5 | 17.9 | BioSimBench 1.0: Claude Haiku 5.5 · CellType |