DDR-Bench - Unsupported Factual Insights - MIMIC: leaderboard
Metric: Unsupported Factual Insight Rate (%). Source: arxiv.org. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 0 |
| 2 | DeepSeek V3.2 | 0 |
| 3 | GLM-4.6 | 0 |
| 4 | Gemini 3 Flash | 0 |
| 5 | GPT-5 Mini | 2.38 |
| 6 | GPT-5.2 | 2.86 |
| 7 | Qwen 3 Next 80B A3B Instruct | 5.26 |
| 8 | Claude Sonnet 4.5 | 6.15 |
| 9 | MiniMax-M2 | 9.09 |
Interactive version: theaggregate.ai/benchmark?slug=ddr-bench-unsupported-factual-insights-mimic · How It Works · Data refreshed daily, snapshot 2026-09-19.