Share your thoughts, 1 month free Claude Pro on usSee more

Deep Search on BrowseComp-ZH

66.6Accuracy

GLM-4.7

Updated 5d ago

Evaluation Results

Method	Links
GLM-4.7 2026.05		66.6
TaS 2026.02		63.7
GPT-5 High-Think 2026.02		63
GPT-5 Medium-Think 2026.02		62.9
OpenAI-o3 2026.05		58.1
GPT-5 Medium-Think 2026.02		56.5
MiroThinker-v1.0-72B 2026.02		55.6
MiroThinker-v1.0-30B 2026.02		47.8
Tongyi DeepResearch (30B) 2026.02		46.7
Qwen3-30B-A3B-thinking-SFT + SAPO 2026.05		45.7
GLM-4.6 2026.05		45.1
Tongyi DeepResearch 2026.05		43.2
OpenAI Deep Research 2026.02		42.9
MiroThinker-v1.0-8B 2026.02		40.2
TaS 2026.02		35.3
TaS 2026.02		34.9
Qwen3-Max 2026.02		34.3
Qwen3-30B-A3B-thinking-SFT + GRPO 2026.05		33.2
Claude-4-Sonnet (Thinking) 2026.02		29.1
Kimi K2 2026.05		28.8
Gemini-2.5-Flash 2026.02		28.4
Gemini-2.5-Pro 2026.02		27.8
Qwen3-30B-A3B-thinking-SFT 2026.05		27.3
Gemini-2.5-Flash 2026.02		26.6
Web-30B-E-GRPO 2026.05		26.4
Qwen3-Max 2026.02		23.5
Qwen3-8B-SFT + SAPO 2026.05		22.2
Qwen3-8B-SFT + GRPO 2026.05		19.7
WebSailor-32B 2026.05		18
Qwen3-8B-SFT 2026.05		17.9
MiroThinker-v1.0-8B 2026.05		16.3
Qwen3-8B-SFT + ARPO 2026.05		15.9
WebSailor-7B 2026.05		9.8
WebExplorer-8B 2026.05		6.2
MiroThinker-v1.5-30B 2026.05		5.7