DDR-Bench - Unsupported Factual Insights - Average: leaderboard
Metric: Unsupported Factual Insight Rate (%). Source: arxiv.org. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash | 0.61 |
| 2 | Gemini 2.5 Pro | 0.85 |
| 3 | GPT-5.2 | 0.95 |
| 4 | GPT-5 Mini | 1.1 |
| 5 | GLM-4.6 | 1.65 |
| 6 | DeepSeek V3.2 | 1.77 |
| 7 | Claude Sonnet 4.5 | 4.08 |
| 8 | MiniMax-M2 | 5.59 |
| 9 | Qwen 3 Next 80B A3B Instruct | 5.69 |
Interactive version: theaggregate.ai/benchmark?slug=ddr-bench-unsupported-factual-insights-average · How It Works · Data refreshed daily, snapshot 2026-09-19.