Benchmarks
357 sources · 1,263 models × 3,079 benchmarks · 162,127 scoresTests whether agents can analyze biological data, laboratory results and patient records to understand how respiratory viruses cause illness, how immunity protects people and whether vaccines and treatments stay effective as the viruses change. 64 evaluations across viral structure and function, immunity and antigenic variation, host response and disease biology, antiviral activity and resistance, and clinical and population outcomes, each pairing a question and a published dataset with task-specific grading criteria; scores are the mean pass rate over three offline runs per evaluation, with refusals and errors counted as failures. benchmarks.bio excludes it from its overall Capabilities score.
186th strongest of 818 fitted benchmarks
Signal-to-noise +10.49
40th hardest of 818 fitted benchmarks
Capability needed ≳ 181.5
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | Gemini 4 Carbon | not reported · pi | 30.2 | |
| 2 | Claude Sonnet 5.5 | not reported · pi | 18.8 | |
| 3 | Claude Fable 5.2 | not reported · Claude Code | 17.7 | |
| 4 | GPT-6 Astra | not reported · pi | 13.5 | |
| 5 | Claude Opus 5.5 | not reported · Claude Code | 13.0 | |
| 6 | Gemini 4 Argon | not reported · pi | 12.0 | |
| 7 | GPT-6.1 Terra | not reported · pi | 11.5 | |
| 8 | Grok 4.7 | not reported · Pi | 11.5 | |
| 9 | GPT-6.1 Sol | not reported · pi | 10.9 | |
| 10 | Grok 4.6 | not reported · pi | 9.4 | |
| 11 | DeepSeek V4.1 Flash | not reported · pi | 6.3 | |
| 12 | GLM-5.3 | not reported · pi | 6.3 | |
| 13 | Mistral Large 4 | not reported · pi | 4.2 | |
| 14 | Gemini 3.8 Flash | not reported · pi | 2.6 | |
| 15 | GPT-6 Luna | not reported · pi | 2.1 | |
| 16 | Kimi K3.1 Pro | not reported · pi | 1.6 | |
| 17 | Nemotron 3.5 Lightning 30B A3B (precision unspecified) | not reported · pi | 0.0 |
Benchmarks outside its family whose model rankings track it most reliably (Spearman ρ over at least 10 shared models).