CHARM - Boundary Awareness: leaderboard
Character hallucination in multicultural role play: whether the model recognises that a question crosses the character's knowledge boundary. Accuracy, higher is better. Source: arxiv.org/abs/2609.01352.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 8B | 94.1 | #1130 |
| 2 | Gemini 3.5 Flash | 94 | #66 |
| 3 | GPT-4o | 91.3 | #357 |
| 4 | Gemma 3 12B | 87.8 | #674 |
| 5 | GPT-5.5 | 86.9 | #33 |
| 6 | Qwen 3 8B | 62.6 | #678 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=charm-boundary-awareness · How It Works · Data refreshed daily, snapshot 2026-11-05.