Multi-step navigation and information location on BrowseComp-ZH
68.7ScoreOpenAI-o3
Evaluation results
| Method | Links | |
|---|---|---|
| OpenAI-o3# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 68.7 | |
| GLM-4.7-357B# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 66.6 | |
| DeepSeek-V3.2-671B# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 65 | |
| GPT-5-High# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 63 | |
| GLM-4.6-357B# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 49.5 | |
| DeepSeek-V3.1-671B# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 49.2 | |
| Minimax-M2-230B# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 48.5 | |
| OpenSeeker-v1-30B-SFT# Samples=11.7 k, # OS Samples=11.7 k, Training=SFT, Academic=true2026.03 | 48.4 | |
| Tongyi DeepResearch# Samples=?, # OS Samples=0, Training=CPT + SFT + RL, Academic=false2026.03 | 46.7 | |
| WebSailor-V2-30B# Samples=?, # OS Samples=0, Training=SFT + RL, Academic=false2026.03 | 44.1 | |
| OpenAI Deep Research# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 42.9 | |
| Claude-4.5-Sonnet# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 42.4 | |
| Claude-4-Opus# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 37.4 | |
| DeepDive-32B# Samples=4.1 k, # OS Samples=4.1 k, Training=SFT + RL, Academic=false2026.03 | 29.7 | |
| Kimi-K2-Instruct-1T# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 28.8 | |
| WebSailor-V2-30B# Samples=?, # OS Samples=0, Training=SFT, Academic=false2026.03 | 28.3 | |
| WebSailor-32B# Samples=?, # OS Samples=0, Training=SFT + RL, Academic=false2026.03 | 25.5 | |
| Claude-4-Sonnet# Samples=?, # OS Samples=0, Training=?, Academic=false2026.03 | 22.5 | |
| WebDancer-32B# Samples=?, # OS Samples=0, Training=SFT + RL, Academic=false2026.03 | 18 | |
| MiroThinker-32B-v0.1# Samples=147 k, # OS Samples=147 k, Training=SFT + RL, Academic=false2026.03 | 17 | |
| MiroThinker-32B-v0.1# Samples=147 k, # OS Samples=147 k, Training=SFT, Academic=false2026.03 | 13.8 |