EduResearchBench - Policy Analysis: leaderboard
Metric: Judge score (1-5; 4 atomic tasks, Chinese and English, mean of Doubao-seed-1.6 and DeepSeek-V3.2 critique-then-score ratings). Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 3.58 |
| 2 | Gemini 3 Flash | 3.36 |
| 3 | Qwen 2.5 72B Instruct | 2.62 |
| 4 | Llama 3 70B Instruct | 2.34 |
| 5 | LongWriter-glm4-9B | 2.26 |
Interactive version: theaggregate.ai/benchmark?slug=eduresearchbench-policy-analysis · How It Works · Data refreshed daily, snapshot 2026-09-25.