Pressure Reveals Character - Honesty: Honesty: leaderboard
Mean 1-5 judge rubric score over 44 multi-turn scenarios testing giving truthful information. One behaviour of the Honesty category; a subset of the headline's transcripts.
Metric: Mean judge rubric score over 44 scenarios (1 = failure, 5 = exemplary). Source: storage.googleapis.com. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 | 4.68 | #156 |
| 2 | Claude Sonnet 4.5 | 4.57 | #250 |
| 3 | GPT-5.2 | 4.43 | #176 |
| 4 | Claude Sonnet 4 | 4.41 | #400 |
| 5 | GPT-5 Mini | 4.11 | #348 |
| 6 | Gemini 3 Pro | 3.84 | #137 |
| 7 | Llama 4 Maverick | 3.84 | #1099 |
| 8 | GPT-4o | 3.7 | #741 |
| 9 | Llama 3.1 405B Instruct | 3.55 | #1115 |
| 10 | Kimi K2 | 3.52 | #439 |
| 11 | Grok 4.1 | 3.5 | #476 |
| 12 | Gemini 2.5 Flash | 3.48 | #453 |
| 13 | Llama 3.3 70B | 3.36 | #1498 |
| 14 | Qwen 3 235B A22B | 3.25 | #623 |
| 15 | Gemini 2.5 Pro | 3.23 | #266 |
Interactive version: theaggregate.ai/benchmark?slug=pressure-reveals-character-honesty-honesty · How It Works · Data refreshed daily, snapshot 2026-11-13.