HELM AIR-Bench 2024 - #14.26: Health conditions: leaderboard
HELM AIR-Bench 2024 (release v1.19.0) refusal rate for the level-4 risk category "#14.26: Health conditions", under level-3 "#14: Hate Speech (Inciting/Promoting/Expressing hatred)" and level-2 "Hate/Toxicity". The category holds only 15 prompts, so scores move in coarse steps. A GPT-4o judge grades each response from compliance (0) to refusal (1); the mean is shown as 0-100, higher meaning more refusals (safer).
Metric: Refusal Rate (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 87 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 100 | #105 |
| 2 | Claude Sonnet 4.5 | 100 | #148 |
| 3 | GPT-OSS-120B | 100 | #346 |
| 4 | Claude 3.5 Sonnet (20241022) | 100 | #302 |
| 5 | GPT-5 Nano | 100 | #439 |
| 6 | Gemini 3 Pro (Preview) | 100 | #71 |
| 7 | Gemini 1.5 Pro (002) | 100 | #368 |
| 8 | Gemini 1.5 Flash (002) | 100 | #587 |
| 9 | Claude 3.5 Sonnet (20240620) | 100 | #308 |
| 10 | GPT-OSS-20B | 100 | #523 |
| 11 | GPT-4.1 (2025-04-14) | 100 | #257 |
| 12 | GPT-4o (2024-08-06) | 100 | #343 |
| 13 | Llama 3 8B Instruct | 100 | #1116 |
| 14 | Gemini 2.5 Flash Lite | 100 | #432 |
| 15 | Qwen 3 Next 80B A3B (Thinking) | 100 | #312 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-air-bench-2024-14-26-health-conditions · How It Works · Data refreshed daily, snapshot 2026-11-05.