Agentic Workflow Performance Prediction on FLORA-Bench Reason-GD 1.0 (test)
73.8AccuracyGLOW
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GLOWbase_LLM=Qwen3-1.7B, GNN=two-layer GAT, SBERT=all-MiniLM-L6-v22025.12 | 73.8 | 66.1 | |
| Agentic Predictor2025.12 | 72.6 | 63.1 | |
| Qwen3base_model=Qwen3-1.7B, fine-tuned=true2025.12 | 71.8 | 62.6 | |
| GCNII2025.12 | 71.7 | 62.1 | |
| GCN2025.12 | 71.6 | 62 | |
| Graph Transformer2025.12 | 71.6 | 62.7 | |
| GAT2025.12 | 71.1 | 62.4 | |
| One-For-All2025.12 | 70.9 | 62.7 |