HalluWorld: leaderboard
Hallucination remains a central failure mode of large language models, but existing benchmarks operationalize it inconsistently across summarization, question answering, retrieval-augmented generation, and agentic interaction. This fragmentation makes it unclear whether a mitigation that works in one setting reduces hallucinations across contexts. Current benchmarks either require human annotation
Metric: Overall (self-reported). Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 5 | #31 |
| 2 | O3 | 6.8 | #124 |
| 3 | O4 Mini | 7.4 | #180 |
| 4 | O3 Mini | 8.1 | #272 |
| 5 | Kimi K2.6 | 9.3 | #104 |
| 6 | GPT-4o | 12.6 | #343 |
| 7 | GLM-5 | 14.3 | #141 |
| 8 | DeepSeek V3 (0324) | 15.4 | #340 |
| 9 | GPT-5.4 Mini | 21.5 | #212 |
| 10 | Qwen 3 30B A3B | 23.3 | #493 |
| 11 | GPT-4o Mini | 28.1 | #599 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=halluworld · How It Works · Data refreshed daily, snapshot 2026-10-20.