Web Agent Task Success on WebArena Gitlab 180 tasks
46.7Task Success Rate (SR)ReasoningBank + MaTTS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ReasoningBank + MaTTSBackbone LLM=Gemini-2.5-pro2025.09 | 46.7 | 9.1 | |
| ReasoningBankBackbone LLM=Gemini-2.5-pro2025.09 | 44.4 | 9.8 | |
| ReasoningBank + MaTTSBackbone LLM=Claude-3.7-sonnet2025.09 | 43.3 | 9.4 | |
| ReasoningBank + MaTTSBackbone LLM=Gemini-2.5-flash2025.09 | 42.8 | 11.9 | |
| ReasoningBankBackbone LLM=Claude-3.7-sonnet2025.09 | 41.1 | 9.5 | |
| ReasoningBankBackbone LLM=Gemini-2.5-flash2025.09 | 40.6 | 12.3 | |
| AWMBackbone LLM=Gemini-2.5-pro2025.09 | 40 | 11.2 | |
| SynapseBackbone LLM=Gemini-2.5-pro2025.09 | 38.3 | 11.3 | |
| SynapseBackbone LLM=Claude-3.7-sonnet2025.09 | 38 | 10 | |
| AWMBackbone LLM=Gemini-2.5-flash2025.09 | 37.2 | 13.2 | |
| No MemoryBackbone LLM=Claude-3.7-sonnet2025.09 | 36.7 | 10.6 | |
| SynapseBackbone LLM=Gemini-2.5-flash2025.09 | 35.6 | 13 | |
| No MemoryBackbone LLM=Gemini-2.5-pro2025.09 | 35 | 11.6 | |
| AWMBackbone LLM=Claude-3.7-sonnet2025.09 | 34.6 | 10.9 | |
| No MemoryBackbone LLM=Gemini-2.5-flash2025.09 | 33.9 | 13.3 |