EduResearchBench: leaderboard

Metric: Judge score (1-5; overall over 24 atomic research-writing tasks in six modules, 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

#ModelScore
1GPT-53.63
2Gemini 3 Flash3.34
3Qwen 2.5 72B Instruct2.9
4Llama 3 70B Instruct2.46
5LongWriter-glm4-9B2.34

Interactive version: theaggregate.ai/benchmark?slug=eduresearchbench · How It Works · Data refreshed daily, snapshot 2026-09-25.