X-Risks Leaderboard: leaderboard
Extreme-risks evaluation leaderboard for frontier models, using 3C3H scoring over biology, chemistry, and cybersecurity risk-domain questions.
Metric: 3C3H Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 (2024-12-17) | 29.09 | #149 |
| 2 | O3 Mini (2025-01-31) | 27.73 | #240 |
| 3 | O1 Mini (2024-09-12) | 20.78 | #403 |
| 4 | GPT-4o (2024-08-06) | 18.92 | #336 |
| 5 | Llama 3.3 70B Instruct | 17.83 | #534 |
| 6 | Qwen 2.5 72B Instruct | 16.6 | #454 |
| 7 | GPT-4o Mini (2024-07-18) | 15.51 | #558 |
| 8 | Claude 3.5 Sonnet (20241022) | 14.45 | #291 |
| 9 | Claude 3 Haiku (20240307) | 13.06 | #771 |
| 10 | QwQ 32B-Preview | 11.03 | #651 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=x-risks-leaderboard · How It Works · Data refreshed daily, snapshot 2026-10-20.