Delulu - Import Hallucination Detection: leaderboard
LLMs as judges on the 478 Delulu (2026) fill-in-the-middle samples whose hallucinated completion contains imports of packages that do not exist: the judge must accept the real completion and reject the hallucinated one. The score is the both-correct rate (%); higher is better.
Metric: Both-correct rate (%) on the 478 Delulu samples whose hallucination is imports of packages that do not exist, as an LLM judge: the judge sees each sample twice, once with the gold completion and once with the hallucinated one, and must accept the first and reject the second (both-correct rate, %), with chain-of-thought and a binary verdict; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 | 83.4 | #165 |
| 2 | Claude Sonnet 4.6 | 80.3 | #179 |
| 3 | GPT-5.2 Codex | 75.4 | #186 |
| 4 | GPT-5.4 Mini | 69.2 | #409 |
| 5 | GPT-5.4 | 67.9 | #138 |
| 6 | Claude Haiku 4.5 | 65.6 | #587 |
| 7 | GPT-4.1 Mini | 56.9 | #799 |
| 8 | GPT-4o Mini | 48.7 | #1680 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=delulu-import-hallucination-detection · How It Works · Data refreshed daily, snapshot 2026-11-22.