Deep Search on Browse Comp-ZH
81.3ScoreSeed1.8
Evaluation Results
| Method | Links | |
|---|---|---|
| Seed1.8Size=-, Category=Commercial Closed-source Models2026.04 | 81.3 | |
| GPT-5.2 (xhigh)Size=-, Category=Commercial Closed-source Models2026.04 | 76.1 | |
| GLM-5Size=744B-A40B, Category=Open-source Large-scale Models (w/o ctm)2026.04 | 75.9 | |
| MiroThinker-v1.7-miniModel Category=Trained Agents (≥30B)2026.03 | 72.3 | |
| MiroThinker-v1.5-235BModel Category=Trained Agents (≥30B)2026.03 | 71.5 | |
| GLM-4.7Size=355B-A32B, Category=Open-source Large-scale Models (w/o ctm)2026.04 | 67.5 | |
| Gemini-3.0-ProModel Category=Foundation Models with Tools2026.03 | 66.8 | |
| MiroThinker-v1.5-30BModel Category=Trained Agents (≥30B)2026.03 | 66.8 | |
| Gemini-3-ProSize=-, Category=Commercial Closed-source Models2026.04 | 66.8 | |
| GLM-4.7Model Category=Foundation Models with Tools2026.03 | 66.6 | |
| DeepSeek-V3.2-671BCategory=Large Foundation Model2026.02 | 65 | |
| DeepSeek-V3.2Model Category=Foundation Models with Tools2026.03 | 65 | |
| OpenAI GPT-5 HighModel Category=Foundation Models with Tools2026.03 | 65 | |
| DeepSeek-V3.2Size=671B-A37B, Category=Open-source Large-scale Models (w/o ctm)2026.04 | 65 | |
| Claude-4.5-OpusModel Category=Foundation Models with Tools2026.03 | 62.4 | |
| Kimi-K2.5Model Category=Foundation Models with Tools2026.03 | 62.3 | |
| Claude Opus 4.5Size=-, Category=Commercial Closed-source Models2026.04 | 62.3 | |
| OpenAI-o3Model Category=Foundation Models with Tools2026.03 | 58.1 | |
| QuarkMedSearchSize=30B-A3B, Category=Ours2026.04 | 57.55 | |
| MiroThinker-v1.0-72BModel Category=Trained Agents (≥30B)2026.03 | 55.6 | |
| REDSearcherSize=30B-A3B, Category=Open-source Medium-scale Models (w/o ctm)2026.04 | 49.8 | |
| GLM-4.6-357BCategory=Large Foundation Model2026.02 | 49.5 | |
| Minimax-M2-230BCategory=Large Foundation Model2026.02 | 48.5 | |
| OpenSeeker-30B-SFTModel Category=Trained Agents (≥30B)2026.03 | 48.4 | |
| Minimax-M2.1Model Category=Foundation Models with Tools2026.03 | 47.8 | |
| MiroThinker-v1.0-30BModel Category=Trained Agents (≥30B)2026.03 | 47.8 | |
| Marco-DR-8BModel Category=Trained Agents (≤8B)2026.03 | 47.1 | |
| Tongyi-DeepResearch-30BCategory=Research Agent2026.02 | 46.7 | |
| Tongyi-DR-30BModel Category=Trained Agents (≥30B)2026.03 | 46.7 | |
| Tongyi DeepResearchSize=30B-A3B, Category=Open-source Medium-scale Models (w/o ctm)2026.04 | 46.7 | |
| Base (Tongyi DeepResearch)†Size=30B-A3B, Category=Ours, Evaluation Environment=Unified2026.04 | 45.4 | |
| WebResearcherSize=30B-A3B, Category=Open-source Medium-scale Models (w/o ctm)2026.04 | 45.2 | |
| WebSailor-V2-30BModel Category=Trained Agents (≥30B)2026.03 | 44.1 | |
| WebSailorV2Size=30B-A3B, Category=Open-source Medium-scale Models (w/o ctm)2026.04 | 44.1 | |
| MiroThinker-v1.0-8BCategory=Research Agent2026.02 | 40.2 | |
| MiroThinker-v1.0-8BModel Category=Trained Agents (≤8B)2026.03 | 40.2 | |
| DeepMiner-32B-RLModel Category=Trained Agents (≥30B)2026.03 | 40.1 | |
| RE-TRAC-4BModel Category=Trained Agents (≤8B)2026.03 | 36.1 | |
| WebExplorer-8B-RLModel Category=Trained Agents (≤8B)2026.03 | 32 | |
| Nanbeige4.1-3BCategory=Ours2026.02 | 31.83 | |
| Claude-4-SonnetModel Category=Foundation Models with Tools2026.03 | 29.1 | |
| AgentCPM-Explore-4BModel Category=Trained Agents (≤8B)2026.03 | 29.1 | |
| AgentCPM-Explore-4BCategory=Research Agent2026.02 | 29 | |
| Qwen3-Next-80B-A3BCategory=Small Foundation Model2026.02 | 8.25 | |
| Qwen3-4B-2507Category=Small Foundation Model2026.02 | 7.92 | |
| Qwen3-32BCategory=Small Foundation Model2026.02 | 7.34 | |
| Qwen3-14BCategory=Small Foundation Model2026.02 | 7.11 | |
| Qwen3-8BCategory=Small Foundation Model2026.02 | 5.15 | |
| Qwen3-30B-A3B-2507Category=Small Foundation Model2026.02 | 4.12 | |
| Nanbeige4-3B-2511Category=Baseline2026.02 | 3.09 |