Pressure Reveals Character - Honesty: Evaluation Awareness: leaderboard
Mean 1-5 judge rubric score over 10 multi-turn scenarios testing behaving the same when it seems to be tested. One behaviour of the Honesty category; a subset of the headline's transcripts.
Metric: Mean judge rubric score over 10 scenarios (1 = failure, 5 = exemplary). Source: storage.googleapis.com. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 5 | #250 |
| 2 | GPT-5.2 | 5 | #176 |
| 3 | Gemini 3 Flash | 5 | #164 |
| 4 | DeepSeek V3.2 | 5 | #406 |
| 5 | GPT-5 Mini | 5 | #348 |
| 6 | Qwen 3 32B | 5 | #960 |
| 7 | Gemini 3 Pro | 5 | #137 |
| 8 | Claude Sonnet 4 | 5 | #400 |
| 9 | Qwen 3 235B A22B | 5 | #623 |
| 10 | GPT-4o | 4.9 | #741 |
| 11 | Gemini 2.5 Flash | 4.9 | #453 |
| 12 | Kimi K2 | 4.9 | #439 |
| 13 | Claude Opus 4.5 | 4.9 | #156 |
| 14 | Mistral Small 3.2 | 4.9 | #1591 |
| 15 | Mistral Large 3 | 4.9 | #817 |
Interactive version: theaggregate.ai/benchmark?slug=pressure-reveals-character-honesty-evaluation-awareness · How It Works · Data refreshed daily, snapshot 2026-11-13.