Web Navigation on WebArena
61.7Overall Success RateIBM CUGA
Evaluation Results
| Method | Links | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IBM CUGAModel=-, Access=Proprietary2026.06 | 61.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OpenAI CUAModel=-, Access=Proprietary2026.06 | 58.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| WebChallengerModel=GLM-4-32B + Qwen2.5-VL-7B, Access=Open-Source, Evaluation Strategy=zero-shot2026.06 | 56.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Vanilla-IBLLM Backbone=Gemini 1.5 Flash2026.06 | 55.68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 99 | — | — | — | |
| ScribeAgentModel=GPT-4o + Qwen2.5-32B, Access=Proprietary2026.06 | 53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ASILLM Backbone=Gemini 1.5 Flash2026.06 | 52.75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 139.4 | — | — | — | |
| STRUCTUREDAGENTBackbone=Claude 3.5 Sonnet2026.03 | 52.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 45 | 52.8 | 72.6 | 45.8 | — | — | — | — | — | — | — | — | — | — | — | |
| Gemini 3 ProCategory=Proprietary, Teacher for A3-SYNTH=true2026.04 | 51.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Vanilla-IBLLM Backbone=Qwen 2.5-72B2026.06 | 50.73 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 130.3 | — | — | — | |
| WALTModel=GPT-5, Access=Proprietary2026.06 | 50.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ASILLM Backbone=Qwen 2.5-72B2026.06 | 49.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 182.2 | — | — | — | |
| ReasoningBankLLM Backbone=Gemini 1.5 Flash2026.06 | 48.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 124.6 | — | — | — | |
| AgentSymbioticModel=Claude 3.5 + Llama3 8B, Access=Proprietary2026.06 | 48.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Mobile-Agent-v3.5Model=Qwen3-VL-32B, Access=Open-Source, Evaluation Strategy=fine-tuned2026.06 | 48.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Vanilla-IBLLM Backbone=Gemini 1.5 Flash2026.06 | 47.77 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 45.7 | — | — | — | |
| ReasoningBankLLM Backbone=Qwen 2.5-72B2026.06 | 47.62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 156.9 | — | — | — | |
| Vanilla-IBLLM Backbone=Gemini 1.5 Flash2026.06 | 47.48 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 71.7 | — | — | — | |
| AWMLLM Backbone=Gemini 1.5 Flash2026.06 | 47.48 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 87.2 | — | — | — | |
| AWMLLM Backbone=Gemini 1.5 Flash2026.06 | 47.43 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 142.7 | — | — | — | |
| AgentOccamBackbone=Claude 3.5 Sonnet2026.03 | 46.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 46.3 | 38 | 60.1 | 46.7 | — | — | — | — | — | — | — | — | — | — | — | |
| AWMLLM Backbone=Qwen 2.5-72B2026.06 | 46.15 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 178.5 | — | — | — | |
| Vanilla-IBLLM Backbone=Qwen 2.5-72B2026.06 | 45.99 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 58.8 | — | — | — | |
| AgentOccam-JudgeModel=GPT-4, Access=Proprietary2026.06 | 45.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AgentOccam-JudgeModel=GPT-4-Turbo2025.10 | 45.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 52.3 | 43.3 | 67 | 38.9 | — | — | — | — | — | — | — | — | 46.2 | 16.7 | — | |
| ReasoningBankLLM Backbone=Gemini 1.5 Flash2026.06 | 45.45 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 54.6 | — | — | — | |
| ASILLM Backbone=Gemini 1.5 Flash2026.06 | 44.97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 94.6 | — | — | — | |
| ASILLM Backbone=Gemini 1.5 Flash2026.06 | 44.74 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 82.8 | — | — | — | |
| Vanilla-IBLLM Backbone=Qwen 2.5-72B2026.06 | 44.34 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 89.7 | — | — | — | |
| ASILLM Backbone=Qwen 2.5-72B2026.06 | 43.67 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 78.8 | — | — | — | |
| AWMLLM Backbone=Qwen 2.5-72B2026.06 | 43.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 103.6 | — | — | — | |
| ASILLM Backbone=Qwen 2.5-72B2026.06 | 43.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 120.5 | — | — | — | |
| Gemini 3.1 Flash L.Category=Proprietary2026.04 | 42.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3.5-27BCategory=Open-weight (base)2026.04 | 41.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| A3-Qwen3.5-9BCategory=A3 fine-tuned, Teacher=Gemini 3 Pro2026.04 | 41.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Agent-as-AnnotatorsModel=Qwen3.5-9B, Access=Open-Source, Evaluation Strategy=fine-tuned2026.06 | 41.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AWMLLM Backbone=Qwen 2.5-72B2026.06 | 41.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 71.8 | — | — | — | |
| AgentOccamBackbone=GPT 4-o2026.03 | 41.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 44.6 | 35.2 | 62.3 | 32.9 | — | — | — | — | — | — | — | — | — | — | — | |
| ActiveContextTaskExecutor=Gemini-3.0-flash2026.04 | 41.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 33.3 | 26.1 | 47.8 | 66.7 | — | 39 | 41.7 | 50.5 | 53.8 | 43.1 | 30.9 | 43.3 | — | — | — | |
| ReasoningBankLLM Backbone=Qwen 2.5-72B2026.06 | 41.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 99.1 | — | — | — | |
| AWMLLM Backbone=Gemini 1.5 Flash2026.06 | 41.18 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 70.8 | — | — | — | |
| ReasoningBankLLM Backbone=Gemini 1.5 Flash2026.06 | 39.93 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 76.8 | — | — | — | |
| Branch-and-Browse + StePModel=GPT-4o2025.10 | 39.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 47.7 | 41.2 | 54.7 | 37.8 | — | — | — | — | — | — | — | — | 32.4 | 18.8 | — | |
| ReasoningBankLLM Backbone=Qwen 2.5-72B2026.06 | 39.75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 65.4 | — | — | — | |
| API Hybrid AgentModel=GPT-4o2025.10 | 38.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 45.9 | 25.7 | 51.9 | 44.4 | — | — | — | — | — | — | — | — | 41.2 | 16.7 | — | |
| HMTBackbone=GPT-42026.03 | 38.7 | — | — | — | — | — | — | — | — | — | — | — | — | 34.1 | — | 42.2 | 33.8 | 52.5 | 37.6 | 5.2 | — | — | — | — | — | — | — | — | — | — | |
| Vanilla-IBLLM Backbone=GPT-4o-mini2026.06 | 38.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 54.3 | — | — | — | |
| WebPilotModel=GPT-4o, Access=Proprietary2026.06 | 37.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| WebPilotModel=GPT-4o2025.10 | 37.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 33.9 | 36.9 | 65.1 | 39.4 | — | — | — | — | — | — | — | — | 24.7 | — | — | |
| Full ContextTaskExecutor=Gemini-3.0-flash2026.04 | 36.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 10 | 28.3 | 52.2 | 66.7 | — | 42.2 | 33.3 | 58.6 | 58.3 | 35.5 | 40.2 | 47.4 | — | — | — | |
| GenericAgentModel=Claude 3.5 Sonnet, Access=Proprietary2026.06 | 36.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Vanilla-IBLLM Backbone=GPT-4o-mini2026.06 | 35.89 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 124.1 | — | — | — | |
| Branch-and-BrowseModel=GPT-4o2025.10 | 35.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 46.8 | 34.6 | 50.9 | 36.7 | — | — | — | — | — | — | — | — | 26.4 | 18.8 | — | |
| AWMBackbone=GPT-42026.03 | 35.5 | — | — | — | — | — | — | — | — | — | — | — | — | 29.1 | — | 43.3 | 30.8 | 50.9 | 31.8 | 5.9 | — | — | — | — | — | — | — | — | — | — | |
| Agent Workflow MemoryModel=GPT-4o, Access=Proprietary2026.06 | 35.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AWMModel=GPT-42025.10 | 35.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 42.2 | 32.1 | 54.7 | 35 | — | — | — | — | — | — | — | — | 29.1 | 18.8 | — | |
| A3-Qwen3.5-4BCategory=A3 fine-tuned, Teacher=Gemini 3 Pro2026.04 | 35.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SynapseTaskExecutor=Gemini-3.0-flash2026.04 | 35.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 26.7 | 19.6 | 56.5 | 43.3 | — | 44.6 | 41.7 | 71.5 | 70 | 44.4 | 36.6 | 53.6 | — | — | — | |
| Vanilla-IBLLM Backbone=GPT-4o-mini2026.06 | 34.59 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 96.2 | — | — | — | |
| ReasoningBankLLM Backbone=GPT-4o-mini2026.06 | 34.25 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 123.1 | — | — | — | |
| StePModel=GPT-4-Turbo2025.10 | 33.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 35.8 | 33.2 | 52.8 | 26.7 | — | — | — | — | — | — | — | — | 32.4 | 12.5 | — | |
| ASILLM Backbone=GPT-4o-mini2026.06 | 33.15 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 73.4 | — | — | — | |
| StePBackbone=GPT-42026.03 | 33 | — | — | — | — | — | — | — | — | — | — | — | — | 24 | — | 30 | 37 | 59 | 32 | 8.5 | — | — | — | — | — | — | — | — | — | — | |
| ASILLM Backbone=GPT-4o-mini2026.06 | 32.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 138.5 | — | — | — | |
| Training-Free GRPOTaskExecutor=Gemini-3.0-flash2026.04 | 32.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 30 | 17.4 | 56.5 | 36.7 | — | 55.4 | 36.1 | 70.6 | 76.9 | 47.8 | 50.1 | 60.6 | — | — | — | |
| AWMLLM Backbone=GPT-4o-mini2026.06 | 32.24 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 113.4 | — | — | — | |
| StePBackbone=GPT 4-o2026.03 | 32.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 33.3 | 26.9 | 56.4 | 22.5 | — | — | — | — | — | — | — | — | — | — | — | |
| Flat RetrievalBackbone=GPT-42026.03 | 32.1 | — | — | — | — | — | — | — | — | — | — | — | — | 25.8 | — | 39 | 28.5 | 47 | 28.5 | 6.5 | — | — | — | — | — | — | — | — | — | — | |
| ActiveContextTaskExecutor=Gemini-2.5-flash2026.04 | 32.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.3 | 23.9 | 47.8 | 50 | — | 28 | 33.3 | 51.6 | 53.1 | 34.3 | 28 | 38.6 | — | — | — | |
| GenericAgentModel=GPT-4o, Access=Proprietary2026.06 | 31.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BasicClaudeActionBackbone=Claude 3.5 Sonnet2026.03 | 31 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 37 | 30.3 | 30.3 | 28.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3.5-9BCategory=Open-weight (base)2026.04 | 31 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Full ContextTaskExecutor=Gemini-2.5-flash2026.04 | 30.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.3 | 21.7 | 39.1 | 50 | — | 31.9 | 33.3 | 52.7 | 66.5 | 32.6 | 30.4 | 42.6 | — | — | — | |
| ActiveContextTaskExecutor=gpt-4o2026.04 | 30.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.3 | 19.6 | 52.2 | 56.7 | — | 40.6 | 22.2 | 46.4 | 54.7 | 33.8 | 25.9 | 40.8 | — | — | — | |
| AWMLLM Backbone=GPT-4o-mini2026.06 | 29.95 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 62 | — | — | — | |
| SkillWeaverModel=GPT-4o, Access=Proprietary2026.06 | 29.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AWMLLM Backbone=GPT-4o-mini2026.06 | 29.56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 80.4 | — | — | — | |
| SynapseTaskExecutor=Gemini-2.5-flash2026.04 | 29.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.3 | 13 | 47.8 | 40 | — | 27.2 | 41.7 | 54 | 70.2 | 33.3 | 33.5 | 42.9 | — | — | — | |
| SKILL-DISCO + ReActBackbone=GPT-4o2026.06 | 29.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.11 | — | — | — | — | — | — | 51.068 | — | — | 0.1215 | |
| ASILLM Backbone=GPT-4o-mini2026.06 | 28.93 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 87.2 | — | — | — | |
| WebarenaReplicationBackbone=Claude 3.5 Sonnet2026.03 | 28.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 31.5 | 31.3 | 19.7 | 28.8 | — | — | — | — | — | — | — | — | — | — | — | |
| Training-Free GRPOTaskExecutor=Gemini-2.5-flash2026.04 | 27.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 3.3 | 21.7 | 52.2 | 33.3 | — | 51.7 | 36.1 | 62.3 | 80.6 | 43.7 | 64.4 | 60.5 | — | — | — | |
| Full ContextTaskExecutor=gpt-4o2026.04 | 27.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.3 | 13 | 43.5 | 53.3 | — | 38.9 | 25 | 46.4 | 54.1 | 27.5 | 37.5 | 41.5 | — | — | — | |
| AutoEvalModel=GPT-42025.10 | 26.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 27.5 | 39.6 | 20.8 | 25 | — | — | — | — | — | — | — | — | 20.9 | 16.7 | — | |
| ReasoningBankLLM Backbone=GPT-4o-mini2026.06 | 26.38 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 57.9 | — | — | — | |
| TTIModel=Gemma 3 12B, Access=Open-Source, Evaluation Strategy=fine-tuned2026.06 | 26.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GPT-4.1 baselineBase model=GPT-4.12026.04 | 25.77 | — | — | — | — | — | — | — | — | — | — | — | — | 24.56 | — | — | 30.91 | 15.38 | 26.79 | — | — | — | — | — | — | — | — | — | — | — | |
| Training-Free GRPOTaskExecutor=gpt-4o2026.04 | 25.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 23.3 | 10.9 | 43.5 | 33.3 | — | 50.8 | 27.8 | 53.8 | 64.8 | 41.4 | 31.3 | 49.1 | — | — | — | |
| No MemoryTaskExecutor=gpt-4o2026.04 | 24.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 6.7 | 13 | 47.8 | 33.3 | — | 49.5 | 33.3 | 63.9 | 73.9 | 35.6 | 35.8 | 52.6 | — | — | — | |
| ASIofflineBackbone=GPT-4o2026.06 | 24.63 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.71 | — | — | — | — | — | — | 56.662 | — | — | 0.1269 | |
| AutoSurfer_webarenaBase model=Qwen2.52026.04 | 24.23 | — | — | — | — | — | — | — | — | — | — | — | — | 24.56 | — | — | 20 | 26.92 | 26.79 | — | — | — | — | — | — | — | — | — | — | — | |
| Learn-by-InteractModel=Codestral 22B, Access=Open-Source, Evaluation Strategy=fine-tuned2026.06 | 24.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3.5-4BCategory=Open-weight (base)2026.04 | 24.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReActBackbone=GPT-4o2026.06 | 23.89 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.88 | — | — | — | — | — | — | 50.914 | — | — | 0.1152 | |
| AutoSurfer_wllmBase model=Qwen2.52026.04 | 23.71 | — | — | — | — | — | — | — | — | — | — | — | — | 22.81 | — | — | 23.64 | 34.62 | 19.64 | — | — | — | — | — | — | — | — | — | — | — | |
| STRACEPrompt Setting=STRACE (Ours)2026.07 | 23.7 | — | — | — | — | — | — | — | — | — | — | — | — | 16.2 | — | — | 30.8 | 36.4 | 17.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Gemini-3.0-flashTaskExecutor=gpt-4o-mini, Memory Management=Zero-Shot Prompting (Curator)2026.04 | 23.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 10 | 17.4 | 30.4 | 40 | — | 37.8 | 25 | 54.6 | 48.1 | 26.1 | 32 | 40.7 | — | — | — | |
| BrowserGymModel=GPT-4o2025.10 | 23.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SynapseTaskExecutor=gpt-4o2026.04 | 23 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 20 | 17.4 | 47.8 | 23.3 | — | 35.2 | 16.7 | 65.6 | 67 | 35.7 | 26.9 | 46.2 | — | — | — | |
| SKILL-DISCO + CodeActBackbone=GPT-4o2026.06 | 22.91 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 8.47 | — | — | — | — | — | — | 85.649 | — | — | 0.1606 | |
| AgentTrekModel=Qwen2.5-32B, Access=Open-Source, Evaluation Strategy=fine-tuned2026.06 | 22.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |