HalluWorld - Grid Compound: leaderboard
Compound hallucination in HalluWorld-Grid (2026), MiniGrid worlds whose ground truth comes from the environment state: the rate of hallucinated answers to compound questions that combine several probe categories. Percent; lower is better.
Metric: Hallucination rate (%) on the cross-category compound probes of HalluWorld-Grid (MiniGrid worlds whose ground truth comes from the environment state), micro-averaged over levels and serializers, 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. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5.5 | Medium | 2.7 | #54 |
| 2 | Claude Opus 4.6 | 4.6 | #113 | |
| 3 | O3 | 5.7 | #237 | |
| 4 | O4 Mini | 5.8 | #368 | |
| 5 | Claude Sonnet 4.6 | Medium | 7.5 | #173 |
| 6 | O3 Mini | 7.6 | #546 | |
| 7 | Kimi K2.6 | Non-reasoning | 8.8 | #192 |
| 8 | GPT-4o | 11.3 | #751 | |
| 9 | DeepSeek V3 (0324) | 12.7 | #715 | |
| 10 | GLM-5 | Non-reasoning | 13.4 | #269 |
| 11 | Qwen 3 30B A3B | Non-reasoning | 23 | #1156 |
| 12 | GPT-5.4 Mini | Non-reasoning | 24 | #406 |
| 13 | GPT-4o Mini | 28.7 | #1667 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=halluworld-grid-compound · How It Works · Data refreshed daily, snapshot 2026-11-21.