TrustGen - Dynamic Fact Checking: leaderboard
TrustGen - Dynamic Fact Checking. TrustGen paper 2502.14296v2 provides native accuracy measurements and exact model versions in Table 5. Internal question answering uses parametric knowledge; fact checking classifies claims against supplied evidence rather than using a live retrieval system. Dynamic QA comes from Wikipedia and dynamic fact checking from fact-checking websites, with varied prompt formats. TrustLLM columns use the paper’s dataset pool and LLM judging against reference answers. HallusionBench and AutoHallusion are the paper’s visual-question evaluations, with their reported subsets kept separate. These source-specific runs are not automatically fused with another publisher’s results. QwQ-32B is held because the table does not distinguish its preview or released checkpoint. Reported metric: Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-3.5 Turbo (0125) | 79.33 | #984 |
| 2 | Yi Lightning | 76.54 | #423 |
| 3 | O1 Mini (2024-09-12) | 76 | #411 |
| 4 | Qwen 2.5 72B Instruct | 75.42 | #463 |
| 5 | Mixtral 8x22B Instruct (v0.1) | 74.86 | #755 |
| 6 | GPT-4o Mini (2024-07-18) | 74.3 | #562 |
| 7 | Claude 3 Haiku (20240307) | 72.63 | #781 |
| 8 | O1 Preview (2024-09-12) | 72.5 | #215 |
| 9 | GLM-4 Plus | 71.51 | #397 |
| 10 | GPT-4o (2024-08-06) | 70.95 | #339 |
| 11 | DeepSeek V2.5 | 69.83 | #490 |
| 12 | Gemini 1.5 Flash (002) | 69.43 | #589 |
| 13 | Claude 3.5 Sonnet (20240620) | 68.72 | #305 |
| 14 | Gemini 1.5 Pro (002) | 68.16 | #361 |
| 15 | Gemma 2 27B (IT) | 62.57 | #653 |
No result here: #5 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=trustgen-dynamic-fact-checking · How It Works · Data refreshed daily, snapshot 2026-10-25.