PhantomBench - Entities (Place): leaderboard
PhantomBench (2026) asks language models about generated terms and entities that do not exist, verified absent from a 2.3-trillion-token corpus, and counts answers that invent content instead of abstaining, as labeled by a Gemini 2.5 Flash judge. This board is the hallucination rate in percent on 1,200 non-existent entities such as events, holidays, elections, disasters and creative works, when the prompt asks where it was discovered or took place, presupposing it exists; lower is better.
Metric: Hallucination rate (%): share of answers that do not abstain on the Phantom-E subset (1,200 generated non-existent entities: events, holidays, elections, disasters, creative works and places), verified absent from the 2.3-trillion-token Dolma corpus, when the prompt asks where the concept was discovered or took place (presupposing it exists); a Gemini 2.5 Flash judge labels abstention; reasoning-model rates count only generations that reached a final answer; lower is better. Source: arxiv.org. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 8B Instruct | 3.67 | #4246 |
| 2 | Qwen 2.5 7B Instruct | 12.25 | #2564 |
| 3 | Qwen 3 14B | 16.08 | #1233 |
| 4 | Qwen 3 8B | 18.59 | #1830 |
| 5 | Gemma 2 9B (IT) | 24.42 | #2262 |
| 6 | Qwen 3 4B | 24.84 | #2281 |
| 7 | Qwen 3 1.7B | 33.25 | #5520 |
| 8 | Mistral 7B Instruct (v0.3) | 35.16 | #6653 |
| 9 | Gemini 2.5 Pro | 36.17 | #266 |
| 10 | Gemini 2.5 Flash | 40.34 | #453 |
| 11 | Qwen 3 32B | 43.09 | #960 |
| 12 | DeepSeek R1 Distill Qwen 32B | 63.95 | #1788 |
| 13 | GPT-OSS-20B (Medium) | 67.53 | #1147 (GPT-OSS-20B) |
| 14 | GPT-OSS-20B (Low) | 80.39 | #1147 (GPT-OSS-20B) |
| 15 | Gemma 3 12B (IT) | 85.17 | #1837 |
Interactive version: theaggregate.ai/benchmark?slug=phantombench-entities-place · How It Works · Data refreshed daily, snapshot 2026-11-13.