DAG Classification and Reconstruction on DAGverse 1.0 (test)
59Number of PredictionsGPT-5.1
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPT-5.12026.03 | 59 | 46 | 78 | 95 | 98 | 41 | 69 | |
| DAGverse-Pipeline2026.03 | 57 | 50 | 88 | 92 | 100 | 50 | 88 | |
| Gemini2.5-Pro2026.03 | 54 | 39 | 72 | 95 | 100 | 39 | 72 | |
| InternVL3.5-30B2026.03 | 37 | 16 | 43 | 63 | 98 | 12 | 32 | |
| InternVL3.5-8B2026.03 | 29 | 9 | 31 | 43 | 93 | 7 | 24 | |
| Qwen3-VL-30B2026.03 | 21 | 12 | 57 | 25 | 97 | 9 | 43 | |
| Qwen3-VL-7B2026.03 | 12 | 6 | 50 | 52 | 96 | 3 | 25 |