HalluWorld - Grid Memory: leaderboard
Memory hallucination in HalluWorld-Grid (2026), MiniGrid worlds whose ground truth comes from the environment state: the rate of hallucinated answers to questions about states seen in earlier observations. Percent; lower is better.
Metric: Hallucination rate (%) on the memory probes (tracking states across earlier observations, 6 levels) 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 | Claude Opus 4.6 | 3.8 | #114 | |
| 2 | O4 Mini | 4.5 | #372 | |
| 3 | O3 Mini | 4.9 | #552 | |
| 4 | Claude Sonnet 4.6 | 5 | #179 | |
| 5 | O3 | 5.5 | #242 | |
| 6 | GPT-5.5 | Medium | 5.9 | #56 |
| 7 | Kimi K2.6 | Non-reasoning | 6.2 | #197 |
| 8 | DeepSeek V3 (0324) | 14.3 | #730 | |
| 9 | GLM-5 | Non-reasoning | 15.8 | #276 |
| 10 | GPT-4o | 17.2 | #754 | |
| 11 | GPT-5.4 Mini | Non-reasoning | 26.6 | #409 |
| 12 | Qwen 3 30B A3B | Non-reasoning | 27.7 | #1165 |
| 13 | GPT-4o Mini | 39 | #1680 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=halluworld-grid-memory · How It Works · Data refreshed daily, snapshot 2026-11-22.