List your company

Deep Research Instruction Following on DeepResearchIF

56.4Query-Level AccuracyGPT-5.2
2.00816.12930.2544.371Jun 13, 2026

Evaluation results

MethodLinks
GPT-5.22026.0656.489.271.6
Claude-4.6-Sonnet-Thinking2026.0633.681.159.6
DeepSeek-V3.2Size=671B2026.0625.87654.8
S1-DeepResearchSize=32B, Train=SFT2026.0625.274.354.2
Gemini-3.1-Pro-Preview2026.0624.871.356.2
Doubao-2.0-Pro2026.0623.477.755.4
Kimi-K2.5Size=1T2026.0618.974.351.2
Qwen3.5-397BSize=397B2026.0614.668.450.4
MiniMax-M2.7Size=230B2026.061163.945.1
GLM-5Size=744B2026.0610.862.747.9
Qwen3-235BSize=235B2026.067.454.943.6
Qwen3-32BSize=32B2026.064.128.940.6