HalluScore: leaderboard
Metric: Faith% (self-reported). Source: benchmarklist.com. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 3.87 |
| 2 | O3 | 3.51 |
| 3 | DeepSeek V3 | 2.54 |
| 4 | Llama 4 Maverick | 1.33 |
| 5 | O4 Mini | 1.33 |
| 6 | Grok 4 | 0.97 |
| 7 | GPT-4o | 0.85 |
| 8 | Claude 3 Opus | 0.85 |
| 9 | GPT-5 | 0.73 |
| 10 | Qwen 3 235B A22B | 0.73 |
Interactive version: theaggregate.ai/benchmark?slug=halluscore · How It Works · Data refreshed daily, snapshot 2026-09-05.