HELM AIR-Bench 2024 - #39-40.49: Classification of individuals - Geographic region: leaderboard
Metric: Refusal Rate (%). Source: crfm.stanford.edu. 87 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Nano | 46.67 |
| 2 | GPT-5 | 43.33 |
| 3 | Claude Sonnet 4.5 | 36.67 |
| 4 | Claude Sonnet 4 (20250514) | 35 |
| 5 | Grok 4 | 31.67 |
| 6 | Claude Haiku 4.5 (20251001) | 26.67 |
| 7 | Qwen 3 Next 80B A3B (Thinking) | 25 |
| 8 | Claude 3.5 Sonnet (20241022) | 23.33 |
| 9 | Gemini 1.5 Flash (002) | 23.33 |
| 10 | GPT-5 Mini (2025-08-07) | 21.67 |
| 11 | Gemini 1.5 Pro (002) | 18.33 |
| 12 | O4 Mini (2025-04-16) | 18.33 |
| 13 | Claude Opus 4 (20250514) | 18.33 |
| 14 | Gemini 2.5 Pro (Preview 03-25) | 16.67 |
| 15 | Gemini 3 Pro (Preview) | 16.67 |
Interactive version: theaggregate.ai/benchmark?slug=helm-air-bench-2024-39-40-49-classification-of-individuals-geographic-region · How It Works · Data refreshed daily, snapshot 2026-09-19.