TrustGen - Dynamic QA: leaderboard
TrustGen - Dynamic QA. 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-4o (2024-08-06) | 81.25 | #339 |
| 2 | O1 Preview (2024-09-12) | 79.5 | #215 |
| 3 | Llama 3.1 70B Instruct | 78.12 | #568 |
| 4 | DeepSeek V2.5 | 78.12 | #490 |
| 5 | Claude 3.5 Sonnet (20240620) | 77.08 | #305 |
| 6 | Mixtral 8x22B Instruct (v0.1) | 77.08 | #755 |
| 7 | Yi Lightning | 77.08 | #423 |
| 8 | Qwen 2.5 72B Instruct | 76.04 | #463 |
| 9 | GLM-4 Plus | 76.04 | #397 |
| 10 | O1 Mini (2024-09-12) | 75.5 | #411 |
| 11 | Gemini 1.5 Pro (002) | 75.26 | #361 |
| 12 | GPT-3.5 Turbo (0125) | 75 | #984 |
| 13 | Gemma 2 27B (IT) | 74.48 | #653 |
| 14 | Claude 3 Haiku (20240307) | 73.44 | #781 |
| 15 | Llama 3.1 8B Instruct | 72.4 | #1044 |
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-qa · How It Works · Data refreshed daily, snapshot 2026-10-25.