Query-Item Relevance on RecRM-Bench
89.36AccuracyOurs
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ours2026.05 | 89.36 | 89.12 | |
| GPT-4.12026.05 | 79.26 | 79.48 | |
| Qwen3-MaxThinking=false2026.05 | 76.64 | 77.02 | |
| LongCat-Flash-ThinkingThinking=true2026.05 | 75.97 | 76.17 | |
| Qwen3-MaxThinking=true2026.05 | 75.89 | 76.16 | |
| Deepseek-V3.2Thinking=true2026.05 | 75.22 | 75.57 | |
| Deepseek-V3.2Thinking=false2026.05 | 74.6 | 74.76 | |
| LongCat-Flash-ChatThinking=false2026.05 | 73.18 | 72.82 |