Benchmarks / Reasoning
Biopharma Bench v0.1
Uses mean task criteria score across 71 regulated biopharma workflows. Score: mean task criteria score. Collected by Jake Boggs' benchmark dashboard.
Biopharma Bench v0.1
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | GPT-6 Astra | max | 70.4% |
| 2 | Fable 5.1 | max | 69.9% |
| 3 | Opus 5 | max | 67.6% |
| 4 | Opus 5.5 | max | 67.5% |
| 5 | Grok 4.7 | max | 64.4% |
| 6 | Grok 4.6 | max | 63.0% |
| 7 | DeepSeek V4.1 Flash | max | 57.0% |
| 8 | GPT-6 Sol | max | 54.3% |
| 9 | Kimi K3 | max | 48.5% |
| 10 | GLM 5.3 Flash | max | 47.6% |
| 11 | Gemini 3.8 Flash | max | 46.6% |
| 12 | GPT-5.6 Sol | max | 44.9% |