HELM AIR-Bench 2024 - #8: Celebrating Suffering: leaderboard
HELM AIR-Bench 2024 (release v1.19.0) refusal rate for the level-3 risk category "#8: Celebrating Suffering", under the level-2 category Violence & Extremism of the AIR 2024 regulation-derived taxonomy. Every prompt is a harmful request; a GPT-4o judge grades each response from full compliance (0) to refusal (1), and the mean is shown as 0-100, where higher means 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 | Claude Sonnet 4.5 | 94.2 | #148 |
| 2 | Claude Haiku 4.5 (20251001) | 92.75 | #264 |
| 3 | Claude 3.5 Sonnet (20241022) | 86.23 | #302 |
| 4 | Claude Sonnet 4 (20250514) | 84.78 | #223 |
| 5 | GPT-OSS-120B | 82.61 | #346 |
| 6 | GPT-OSS-20B | 79.71 | #523 |
| 7 | Claude Opus 4 (20250514) | 78.99 | #164 |
| 8 | GPT-5 Mini (2025-08-07) | 74.64 | #175 |
| 9 | GPT-5 | 73.91 | #105 |
| 10 | GPT-5.1 (2025-11-13) | 73.91 | #107 |
| 11 | Qwen 3 Next 80B A3B (Thinking) | 73.19 | #312 |
| 12 | GPT-5 Nano | 71.01 | #439 |
| 13 | Granite 3.3 8B Instruct | 69.57 | #815 |
| 14 | Claude 3.7 Sonnet (20250219) | 65.94 | #209 |
| 15 | Claude 3 Opus (20240229) | 64.49 | #428 |
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-2024-8-celebrating-suffering · How It Works · Data refreshed daily, snapshot 2026-11-05.