List your company

Literature Retrieval on ScholarGym Hard (test)

0.397RecallGPT-5.2
0.034040.128270.22250.31673Jan 29, 2026

Evaluation results

MethodLinks
GPT-5.2Workflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.3970.0450.0810.5050.0080.016
DeepSeek-V3.2Workflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.3720.0170.0320.430.0060.012
Gemini3-ProWorkflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.3650.0280.0520.5040.010.02
DeepSeek-V3.2Workflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=true2026.010.3550.0410.0740.4480.0080.016
Qwen3-30BWorkflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.2260.0160.030.3160.0060.011
GLM-4.7Workflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.2160.0170.0310.3980.0060.012
Qwen3-30BWorkflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=true2026.010.1720.0580.0870.2640.010.019
Qwen3-8BWorkflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=false2026.010.1230.0160.0280.1650.0050.01
Qwen3-8BWorkflow=ScholarGym Agentic, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=5, Extended Thinking=true2026.010.120.0130.0230.2130.0070.014
Qwen3-30BWorkflow=Direct Query Baseline, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=12026.010.0890.0060.0110.0950.0050.01
Qwen3-8BWorkflow=Direct Query Baseline, Retrieval Type=sparse, Content Scope=Abstract-only, Iterations=12026.010.0480.0040.0070.0550.0030.006