Deep Search on BrowseComp v1.0 (test)
55Success RateOpenAI o3*
Evaluation Results
| Method | Links | |
|---|---|---|
| OpenAI o3*2026.06 | 55 | |
| OpenAI DeepResearch*2026.06 | 51.5 | |
| TreeSeekerBackbone=gpt-5.22026.06 | 47 | |
| Flash-SearcherBackbone=gpt-5.22026.06 | 43 | |
| Gemini-DR*2026.06 | 37.8 | |
| IterResearchBackbone=gpt-5.22026.06 | 35.3 | |
| Tongyi-DeepSearch-30B-A3B2026.06 | 33.3 | |
| Perplexity Deep Research*2026.06 | 22 | |
| OpenAI GPT-5*2026.06 | 19.8 | |
| Claude-4.5-Sonnet*2026.06 | 19.6 | |
| LATSBackbone=gpt-5.22026.06 | 16 | |
| Claude-4-Sonnet-Thinking*2026.06 | 14.7 | |
| Grok3 DeepResearch*2026.06 | 12.9 | |
| Kimi K2*2026.06 | 11 | |
| WebSailor-32B*2026.06 | 10.5 | |
| Gemini-2.5-Pro*2026.06 | 9.9 | |
| OpenAI o1*2026.06 | 9.9 | |
| Qwen3-235B-A22B-Instruct-2507*2026.06 | 8 | |
| TreeSeekerBackbone=gpt-4.12026.06 | 7.7 | |
| Flash-SearcherBackbone=gpt-4.12026.06 | 5.7 | |
| WebDancer-32B*2026.06 | 3.8 | |
| Search-o1-32B*2026.06 | 2.8 | |
| DeepSeek-R1*2026.06 | 2 |