Review Quality Evaluation on Scientific Papers 200 sampled papers (random sample)
100Technical DepthGraphReview
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GraphReviewvs Baseline=CNPE-7B2026.05 | 100 | 100 | 100 | 100 | 100 | |
| GraphReviewvs Baseline=DeepReview-7B2026.05 | 99.5 | 99.5 | 99.5 | 99.5 | 99.5 | |
| GraphReviewvs Baseline=DeepReview-14B2026.05 | 99 | 96 | 99 | 98 | 99 | |
| GraphReviewvs Baseline=DeepSeek V3.22026.05 | 98 | 94.5 | 98 | 99 | 98 | |
| GraphReviewvs Baseline=Gemini-2.5-Flash2026.05 | 97.5 | 93.5 | 98 | 98.5 | 98 | |
| GraphReviewvs Baseline=GPT-5-Mini2026.05 | 60 | 55 | 62 | 76 | 62 |