RealityTest: leaderboard
Metric: Text disclosure probability (self-reported). Source: benchmarklist.com. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Haiku 4.5 | 92.3 |
| 2 | GPT-5.1 | 85.6 |
| 3 | Claude Opus 4.6 | 66 |
| 4 | Kimi K2 | 55.2 |
| 5 | Claude Sonnet 4 | 51.5 |
| 6 | Grok 4 | 41.3 |
| 7 | Gemma 4 31B | 31.3 |
| 8 | Gemini 3.1 Pro (Preview) | 30.5 |
| 9 | O4 Mini | 28.3 |
| 10 | DeepSeek R1 | 23.4 |
| 11 | Gemini 3 Flash | 22.7 |
| 12 | DeepSeek V3 | 14.7 |
| 13 | Gemini 2.5 Pro | 12.8 |
| 14 | GPT-4o | 12.6 |
| 15 | Llama 3.3 70B | 12.5 |
Interactive version: theaggregate.ai/benchmark?slug=realitytest · How It Works · Data refreshed daily, snapshot 2026-09-05.