Behavior Prediction on RecRM-Bench
77.78AccuracyOurs
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ours2026.05 | 77.78 | 81.46 | |
| Deepseek-V3.2Thinking=false2026.05 | 47.8 | 52.1 | |
| LongCat-Flash-ChatThinking=false2026.05 | 46.49 | 52.22 | |
| Qwen3-MaxThinking=true2026.05 | 44.47 | 48.5 | |
| Qwen3-MaxThinking=false2026.05 | 42.79 | 52.08 | |
| GPT-4.12026.05 | 39.28 | 49.31 | |
| Deepseek-V3.2Thinking=true2026.05 | 35.17 | 50.76 | |
| LongCat-Flash-ThinkingThinking=true2026.05 | 31.02 | 45.41 |