Pressure Reveals Character - Pass Rate: leaderboard
Share of the 904 pressure scenarios the judge scores 4 or 5 (pass). Same transcripts as the headline mean, counted as pass or not instead of averaged.
Metric: Scenarios scored 4 or 5 of 904 (%). Source: storage.googleapis.com. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 90.04 | #250 |
| 2 | Claude Opus 4.5 | 89.93 | #156 |
| 3 | GPT-5.2 | 87.06 | #176 |
| 4 | Claude Sonnet 4 | 79.87 | #400 |
| 5 | GPT-5 Mini | 79.2 | #348 |
| 6 | Gemini 3 Pro | 70.91 | #137 |
| 7 | GPT-OSS-120B | 66.92 | #700 |
| 8 | Grok 4.1 | 66.7 | #476 |
| 9 | Kimi K2 | 66.59 | #439 |
| 10 | DeepSeek V3.2 | 65.93 | #406 |
| 11 | Qwen 3 235B A22B | 65.6 | #623 |
| 12 | GPT-4o | 64.05 | #741 |
| 13 | Gemini 3 Flash | 63.83 | #164 |
| 14 | Gemini 2.5 Pro | 62.39 | #266 |
| 15 | Gemini 2.5 Flash | 61.06 | #453 |
Interactive version: theaggregate.ai/benchmark?slug=pressure-reveals-character-pass-rate · How It Works · Data refreshed daily, snapshot 2026-11-13.