Interactive Webpage Generation on WebPaper Tasks
93Algorithm: Dynamic Programming (DP) ScorePaperVoyager
Evaluation Results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PaperVoyagerEvaluation Protocol=VLM-Visual, Pipeline Strategy=multi-step scored2026.03 | 93 | 75 | 53.3 | 52.7 | 72 | 82 | 89 | 92.3 | 85 | 93.7 | 92.7 | 95 | 83 | 88 | 89.7 | 85 | 92.3 | 84 | 92 | 84.4 | |
| Claude CodeEvaluation Protocol=rule-based Codegen-Final2026.03 | 56 | 77.9 | 76.5 | 66.2 | 75.7 | 73.9 | 70.7 | 60.5 | 77.7 | 67 | 72.2 | 80.8 | 82.2 | 78.7 | 82 | 55.4 | 82.5 | 66.6 | 76.9 | 72.9 | |
| Kimi-K2Evaluation Protocol=rule-based Codegen-Final2026.03 | 54.8 | 40.6 | 53.9 | 29.8 | 54.7 | 55.3 | 50.4 | 84.3 | 69.1 | 62 | 38.8 | 74.3 | 29.3 | 63.5 | 84.5 | 38.1 | 0 | 62.6 | 52.6 | 54.2 | |
| Gemini-3-ProEvaluation Protocol=rule-based Codegen-Final2026.03 | 51.7 | 80.7 | 55 | 47.3 | 41.7 | 50 | 30.7 | 32 | 55 | 45 | 84 | 51.7 | 84 | 73.3 | 55 | 81.3 | 57.3 | 35 | 85 | 57.5 | |
| GPT5.2Evaluation Protocol=rule-based Codegen-Final2026.03 | 39.9 | 41.7 | 52.6 | 33.9 | 46.3 | 53 | 39.2 | 77.3 | 72.4 | 53.5 | 24.9 | 30.6 | 29.4 | 53.8 | 41.4 | 71.9 | 40.3 | 52.4 | 45.8 | 46.2 | |
| MiniMaxEvaluation Protocol=rule-based Codegen-Final2026.03 | 8.5 | 12.4 | 23.1 | 56.1 | 43.5 | 27.3 | 73.3 | 23.5 | 30.3 | 67.6 | 79 | 83 | 83.5 | 81.3 | 83.9 | 59.5 | 84.8 | 71 | 26.5 | 55 | |
| Qwen-MaxEvaluation Protocol=rule-based Codegen-Final2026.03 | 8 | 12.4 | 42.4 | 28.9 | 29.9 | 27.3 | 69.2 | 27.4 | 44.3 | 79 | 29.5 | 73.5 | 0 | 62.4 | 32.7 | 55.3 | 53.3 | 54.5 | 30.4 | 39.2 |