LLM Evaluation on HealthBench (test)
62.6HealthBench Score (%)GPT-5
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5Rank=1, ∆=→ 02026.03 | 62.6 | 35.4 | |
| GPT-5-miniRank=3, ∆=↓ 12026.03 | 61.7 | 34.8 | |
| DeepSeek-V3.2Rank=4, ∆=↓ 12026.03 | 53 | 31.8 | |
| Gemini 3 FlashRank=5, ∆=→ 02026.03 | 52.5 | 30.4 | |
| Grok-4.1-FastRank=6, ∆=↓ 22026.03 | 52.5 | 30.1 | |
| Qwen3-30BRank=2, ∆=↑ 42026.03 | 49.8 | 34.9 | |
| GPT-4.1Rank=7, ∆=→ 02026.03 | 47 | 23.9 | |
| Claude Sonnet 4.5Rank=8, ∆=→ 02026.03 | 43.5 | 22.7 | |
| GPT-4.1-miniRank=9, ∆=→ 02026.03 | 39.7 | 19.9 | |
| GPT-4.1-nanoRank=10, ∆=→ 02026.03 | 33.9 | 18.1 | |
| Llama-3.1-70BRank=11, ∆=→ 02026.03 | 20.5 | 13.1 |