HalluWorld - Chess (Incorrect FEN): leaderboard
Metric: Hallucination rate (%) over the seven HalluWorld-Chess probes (about 348 questions) when the prompt also includes an incorrect FEN string, temperature 0 where supported, thinking models at medium effort with up to 16k thinking tokens, every model at most 256 answer tokens; lower is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 4 |
| 2 | GPT-5.5 (Medium) | 4 |
| 3 | O3 | 6 |
| 4 | O4 Mini | 8.3 |
| 5 | GPT-5.4 Mini (Medium) | 22.1 |
| 6 | O3 Mini | 24.7 |
| 7 | Claude Opus 4.6 (Medium) | 25.9 |
| 8 | Claude Sonnet 4.6 (Medium) | 26.1 |
| 9 | Claude Opus 4.6 | 33 |
| 10 | Claude Sonnet 4.6 | 39.7 |
| 11 | Kimi K2.6 (Non-reasoning) | 42.2 |
| 12 | GLM-5 (Non-reasoning) | 47.4 |
| 13 | GPT-5.5 (Non-reasoning) | 48.9 |
| 14 | DeepSeek V3.1 (Non-reasoning) | 50 |
| 15 | GPT-5.4 Mini (Non-reasoning) | 54.3 |
Interactive version: theaggregate.ai/benchmark?slug=halluworld-chess-incorrect-fen · How It Works · Data refreshed daily, snapshot 2026-10-07.