Sequential Decision Making on WebShop
0.88ScoreSYMPHONY-L
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SYMPHONY-LModel Size=Stronger2026.01 | 0.88 | 72 | |
| Human ExpertLLM=GPT-42026.01 | 0.82 | 60 | |
| SYMPHONY-SModel Size=Lightweight2026.01 | 0.82 | 56 | |
| MASTERLLM=GPT-42026.01 | 0.8 | — | |
| LATSLLM=GPT-42026.01 | 0.76 | 38 | |
| AgentKitLLM=GPT-42026.01 | 0.7 | — | |
| Fine-tuningLLM=GPT-42026.01 | 0.68 | 45 | |
| ReflexionLLM=GPT-42026.01 | 0.64 | 35 | |
| IL+RLLLM=GPT-42026.01 | 0.62 | 29 | |
| ILLLM=GPT-42026.01 | 0.6 | 29 | |
| ReActLLM=GPT-42026.01 | 0.54 | 32 |