General Agent on BrowseComp w/ Context Manage
67.6ScoreDeepSeek-V3.2 Thinking
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-V3.2 ThinkingThinking Mode=true2026.01 | 67.6 | |
| Kimi-K2 ThinkingThinking Mode=true2026.01 | 60.2 | |
| Gemini-3.0 ProVariant=Pro2026.01 | 59.2 | |
| MiMo-V2-FlashVariant=Flash2026.01 | 58.3 |