Scientific Review Generation on Human Evaluation 50 papers sampled (test)
1,130Actionability (Elo)ProReviewer
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ProReviewer2026.06 | 1,130 | 1,296 | 1,181 | 1,210 | |
| DeepReview2026.06 | 1,018 | 917 | 972 | 953 | |
| AgentReview2026.06 | 974 | 969 | 984 | 984 | |
| AI-Sci-v22026.06 | 967 | 1,035 | 1,046 | 1,034 | |
| CycleReviewer2026.06 | 911 | 784 | 818 | 819 |