PhantomBench - Terms (Date): 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 1,060 non-existent terms modeled on medical abbreviations, science glossaries and legal terms, when the prompt asks when it originated, presupposing it exists; lower is better.
Metric: Hallucination rate (%): share of answers that do not abstain on the Phantom-T subset (1,060 generated non-existent terms modeled on medical abbreviations, science glossaries and legal terms), verified absent from the 2.3-trillion-token Dolma corpus, when the prompt asks when the concept originated or was introduced (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. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 8B Instruct | 0.94 | #4246 |
| 2 | Qwen 3 14B | 1.79 | #1233 |
| 3 | Gemini 2.5 Pro | 2.26 | #266 |
| 4 | Qwen 3 8B | 3.87 | #1830 |
| 5 | Gemma 2 9B (IT) | 4.53 | #2262 |
| 6 | Qwen 2.5 7B Instruct | 5.57 | #2564 |
| 7 | Qwen 3 32B | 6.89 | #960 |
| 8 | Qwen 3 4B | 9.53 | #2281 |
| 9 | Qwen 3 1.7B | 12.64 | #5520 |
| 10 | Gemini 2.5 Flash | 20.85 | #453 |
| 11 | GPT-OSS-20B (Medium) | 30.5 | #1147 (GPT-OSS-20B) |
| 12 | Mistral 7B Instruct (v0.3) | 32.08 | #6653 |
| 13 | DeepSeek R1 Distill Qwen 32B | 43.08 | #1788 |
| 14 | GPT-OSS-20B (Low) | 55.01 | #1147 (GPT-OSS-20B) |
| 15 | Gemma 3 12B (IT) | 69.25 | #1837 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=phantombench-terms-date · How It Works · Data refreshed daily, snapshot 2026-11-13.