List your company
ResearchFormerly WizWand

Textual long-horizon complex reasoning on xBench-DeepSearch

85ScoreDoubao-2.0-Pro
37.1649.586274.42Jun 13, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Doubao-2.0-ProSize=-, Train=-2026.0685
MiniMax-M2.7Size=230B, Train=-2026.0685
Qwen3.5-397BSize=397B, Train=-2026.0684
GPT-5.2Size=-, Train=-2026.0683
GLM-5Size=744B, Train=-2026.0682
Gemini-3.1-Pro-PreviewSize=-, Train=-2026.0680
S1-DeepResearchSize=32B, Train=SFT2026.0679.3
Kimi-K2.5Size=1T, Train=-2026.0678
Claude-4.6-Sonnet-ThinkingSize=-, Train=-2026.0676
Tongyi-DeepResearchSize=30B, Train=Mid/SFT/RL2026.0675
OpenSeeker-v1Size=30B, Train=SFT2026.0674
DeepSeek-V3.2Size=671B, Train=-2026.0671
OpenResearcherSize=30B, Train=SFT2026.0665
Qwen3-235BSize=235B, Train=-2026.0645
Qwen3-32BSize=32B, Train=-2026.0639