Terminal-Bench Science — Life Sciences (Vals)
Science · 2026-09-23
Nineteen life-science tasks under Vals' Terminus 2 protocol with infrastructure/provider-error retries.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 73.7 |
| Claude Opus 5.5 | 52.6 |
| Claude Fable 5 | 21.1 |
| Claude Fable 5.1 | 15.8 |
| Claude Opus 5 | 15.8 |
| Gemini 3.8 Flash | 15.8 |
| GPT-5.6 Sol | 15.8 |
| Muse Spark 1.3 | 15.8 |
| Gemini 3.7 Flash | 10.5 |
| GPT-5.6 Luna | 10.5 |
| Opus 4.8 | 5.3 |
| DeepSeek V4.1 Flash | 5.3 |
| Gemini 3.5 Flash | 5.3 |
| Gemini 3.6 Flash | 5.3 |
| GPT-5.6 Terra | 5.3 |
| Grok 4.6 | 5.3 |
| Kimi K3 | 5.3 |
| MiMo-V2.6-Flash | 5.3 |
| Sonnet 5 | 0.0 |
| DeepSeek-V4-Flash-0731 | 0.0 |
| DeepSeek V4 Pro 0813 | 0.0 |
| GLM-5.3 Flash | 0.0 |
| GLM-5.3 | 0.0 |
| Grok 4.7 | 0.0 |
| Hy4 Preview | 0.0 |
| Mercury 2.5 | 0.0 |
| MiMo-V2.6-Pro | 0.0 |
| Qwen3.8 27B | 0.0 |