Deep Search on BrowseComp-ZH v1.0 (test)
59Success RateOpenAI o3*
Evaluation Results
| Method | Links | |
|---|---|---|
| OpenAI o3*2026.06 | 59 | |
| TreeSeekerBackbone=gpt-5.22026.06 | 43 | |
| OpenAI DeepResearch*2026.06 | 42.9 | |
| Claude-4.5-Sonnet*2026.06 | 40.8 | |
| Flash-SearcherBackbone=gpt-5.22026.06 | 40.3 | |
| OpenAI GPT-5*2026.06 | 34.3 | |
| IterResearchBackbone=gpt-5.22026.06 | 34 | |
| Tongyi-DeepSearch-30B-A3B2026.06 | 33 | |
| Gemini-2.5-Pro*2026.06 | 32.2 | |
| Claude-4-Sonnet-Thinking*2026.06 | 30.8 | |
| OpenAI o1*2026.06 | 29.1 | |
| Doubao DeepResearch*2026.06 | 26 | |
| LATSBackbone=gpt-5.22026.06 | 25.7 | |
| WebSailor-32B*2026.06 | 25.5 | |
| DeepSeek-R1*2026.06 | 23.2 | |
| Qwen3-235B-A22B-Instruct-2507*2026.06 | 23 | |
| Perplexity Deep Research*2026.06 | 22.6 | |
| Kimi K2*2026.06 | 22 | |
| TreeSeekerBackbone=gpt-4.12026.06 | 20.3 | |
| WebDancer-32B*2026.06 | 18 | |
| Search-o1-32B*2026.06 | 17.9 | |
| Flash-SearcherBackbone=gpt-4.12026.06 | 17.7 |