Agentic Web Browsing on BrowseComp
67.6Pass@1DeepSeek-V3.2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-V3.2thinking mode=true, context management technique=true2025.12 | 67.6 | — | |
| Yunque DeepResearchFramework Category=Close-source Agent Frameworks, Model Family=Gemini2026.01 | 62.5 | — | |
| Kimi K2 ThinkingFramework Category=General LLMs + ReAct + Basic Tools2026.01 | 60.2 | — | |
| Kimi-K2thinking mode=true, context management technique=true2025.12 | 60.2 | — | |
| GPT-5 HighFramework Category=General LLMs + ReAct + Basic Tools2026.01 | 54.9 | — | |
| GPT-5 High2025.12 | 54.9 | — | |
| Gemini 3 ProFramework Category=General LLMs + ReAct + Basic Tools2026.01 | 52.5 | — | |
| GLM 4.7Framework Category=General LLMs + ReAct + Basic Tools2026.01 | 52 | — | |
| OpenAI DeepResearchFramework Category=Close-source Agent Frameworks, Model Family=GPT2026.01 | 51.5 | — | |
| DeepSeek-V3.2Framework Category=General LLMs + ReAct + Basic Tools2026.01 | 51.4 | — | |
| DeepSeek-V3.2thinking mode=true, context management technique=false2025.12 | 51.4 | — | |
| OpenAI-o3Paradigm=ReAct, Summary Tool=-2025.09 | 50.9 | — | |
| OpenAI-o3Framework Category=General LLMs + ReAct + Basic Tools2026.01 | 49.7 | — | |
| MiniMax-M2Framework Category=General LLMs + ReAct + Basic Tools2026.01 | 44 | — | |
| MiniMax M22025.12 | 44 | — | |
| Tongyi DeepResearchFramework Category=Open-source Agent Frameworks, Model Family=Qwen2026.01 | 43.4 | — | |
| WebResearcher (w/o TTS)Framework Category=Open-source Agent Frameworks, Model Family=Qwen2026.01 | 37.3 | — | |
| AgentFoldFramework Category=Open-source Agent Frameworks, Model Family=Qwen2026.01 | 36.2 | — | |
| MiroFlowFramework Category=Open-source Agent Frameworks, Model Family=GPT2026.01 | 33.2 | — | |
| DeepSeek-v3.1Paradigm=ReAct, Summary Tool=-2025.09 | 30 | — | |
| Claude-4.5-SonnetFramework Category=General LLMs + ReAct + Basic Tools2026.01 | 24.1 | — | |
| Claude-4.5-Sonnet2025.12 | 24.1 | — | |
| OAgentFramework Category=Open-source Agent Frameworks, Model Family=Claude2026.01 | 22.2 | — | |
| WebSailor-30BParadigm=ReSum, Summary Tool=GPT-OSS-120B2025.09 | 18.8 | 30.9 | |
| WebSailor-30BParadigm=ReSum, Summary Tool=Qwen3-235B2025.09 | 17.2 | 26.7 | |
| WebSailor-30BParadigm=ReSum, Summary Tool=ReSumTool-30B2025.09 | 16 | 25.4 | |
| Kimi-K2Paradigm=ReAct, Summary Tool=-2025.09 | 14.1 | — | |
| WebSailor-30BParadigm=ReSum, Summary Tool=DeepSeek-R1-671B2025.09 | 13.7 | 22.6 | |
| WebSailor-30BParadigm=ReAct, Summary Tool=-2025.09 | 12.8 | 21.8 | |
| WebSailor-30BParadigm=MEM1, Summary Tool=-2025.09 | 12.7 | 22.5 | |
| Claude-4Paradigm=ReAct, Summary Tool=-2025.09 | 12.2 | — | |
| WebSailor-30BParadigm=ReSum, Summary Tool=Qwen3-30B2025.09 | 12.2 | 20.4 | |
| WebSailor-7BParadigm=ReSum, Summary Tool=GPT-OSS-120B2025.09 | 10.5 | 17.2 | |
| WebSailor-7BParadigm=ReSum, Summary Tool=DeepSeek-R1-671B2025.09 | 10.3 | 16.6 | |
| WebSailor-30BParadigm=Recent History, Summary Tool=-2025.09 | 10.3 | 16.7 | |
| WebSailor-7BParadigm=ReSum, Summary Tool=ReSumTool-30B2025.09 | 9 | 15.2 | |
| WebSailor-7BParadigm=ReSum, Summary Tool=Qwen3-235B2025.09 | 8.7 | 15.2 | |
| WebSailor-3BParadigm=ReSum, Summary Tool=GPT-OSS-120B2025.09 | 8.5 | 15.8 | |
| WebSailor-3BParadigm=ReSum, Summary Tool=DeepSeek-R1-671B2025.09 | 7.5 | 13.4 | |
| WebSailor-3BParadigm=ReSum, Summary Tool=ReSumTool-30B2025.09 | 6.8 | 10.8 | |
| WebSailor-7BParadigm=ReSum, Summary Tool=Qwen3-30B2025.09 | 5.8 | 10.3 | |
| WebSailor-3BParadigm=ReSum, Summary Tool=Qwen3-235B2025.09 | 5.7 | 10.3 | |
| WebSailor-7BParadigm=ReAct, Summary Tool=-2025.09 | 5.7 | 10.3 | |
| WebSailor-7BParadigm=Recent History, Summary Tool=-2025.09 | 5.2 | 9.4 | |
| WebSailor-3BParadigm=ReSum, Summary Tool=Qwen3-30B2025.09 | 4.2 | 7.8 | |
| WebSailor-3BParadigm=Recent History, Summary Tool=-2025.09 | 3.8 | 8.9 | |
| WebSailor-3BParadigm=ReAct, Summary Tool=-2025.09 | 3.3 | 5.6 |