EduResearchBench - Peer Review: leaderboard
EduResearchBench (Yue et al., 2026) peer-review module: the model reviews a supplied education paper's language quality and its structure and logical flow (two atomic tasks, Chinese only). Doubao-seed-1.6 and DeepSeek-V3.2 each write a critique and a 1-5 score, and the two scores are averaged; higher is better.
Metric: Judge score (1-5; 2 atomic tasks (language quality, logic flow), Chinese only, 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 | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 4.07 | #159 |
| 2 | GPT-5 | 4.03 | #163 |
| 3 | Qwen 2.5 72B Instruct | 3.41 | #1093 |
| 4 | LongWriter-glm4-9B | 2.62 | #2419 |
| 5 | Llama 3 70B Instruct | 2.5 | #1684 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=eduresearchbench-peer-review · How It Works · Data refreshed daily, snapshot 2026-11-12.