MedHarm - Safe Helpfulness: leaderboard
Metric: Safe helpfulness (%): mean judge score for responses that avoid actionable harm while still giving clinically appropriate general guidance or redirection; MedHarm: 1,100 high-risk medical queries in ten categories of 110 (covert poisoning, lethal dosing, lethal drug interactions, euthanasia, illegal organ harvesting and live anesthesia, drug abuse, medicalized chemical and biological weapons, date-rape drugs, fetal harm and more), all requests a model should decline or redirect; greedy decoding, 512 new tokens; GPT-5-mini judge (91% agreement with human labels on unsafe responses); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.3 Instant | 99.7 |
| 2 | Llama 3.1 8B Instruct | 96.2 |
| 3 | GPT-5.5 | 95.5 |
| 4 | DeepSeek V4 Pro | 92 |
| 5 | Grok 4.3 | 91.8 |
| 6 | Qwen 2.5 7B Instruct | 81.3 |
Interactive version: theaggregate.ai/benchmark?slug=medharm-safe-helpfulness · How It Works · Data refreshed daily, snapshot 2026-09-29.