List your company

Instruction Following on RecRM-Bench

72.66AccuracyOurs
24.830437.247749.66562.0823May 12, 2026

Evaluation results

MethodLinks
Ours2026.0572.6672.4
LongCat-Flash-ChatThinking=false2026.0564.8465.19
GPT-4.12026.0555.4758.33
LongCat-Flash-ThinkingThinking=true2026.0543.7548.16
Qwen3-MaxThinking=false2026.0536.840.33
Deepseek-V3.2Thinking=true2026.0535.2939.5
Deepseek-V3.2Thinking=false2026.0530.4729.94
Qwen3-MaxThinking=true2026.0526.6726.64