Decomposing and Measuring Evaluation Awareness: leaderboard
Metric: HarmBench Saf. (self-reported). Source: benchmarklist.com. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 5.6 |
| 2 | Kimi K2.5 | 4.9 |
| 3 | GPT-OSS-120B | 1.3 |
| 4 | Claude Sonnet 4.5 | 0.5 |
| 5 | GPT-5 | 0.4 |
| 6 | GLM-5 | 0.4 |
| 7 | GPT-OSS-20B | 0.2 |
Interactive version: theaggregate.ai/benchmark?slug=decomposing-and-measuring-evaluation-awareness · How It Works · Data refreshed daily, snapshot 2026-09-05.