Reasoning on ReasonIF
6.7Error Rate (ERR)FLEx
Evaluation Results
| Method | Links | |
|---|---|---|
| FLExBackbone Model=Gemma-27B2026.01 | 6.7 | |
| FLExBackbone Model=Qwen-1.5B2026.01 | 6.7 | |
| FLExBackbone Model=Qwen-3B2026.01 | 9.42 | |
| FLExBackbone Model=Qwen-0.5B2026.01 | 9.77 | |
| FLExBackbone Model=Gemma-1B2026.01 | 22.08 | |
| FLExBackbone Model=Gemma-12B2026.01 | 27.1 | |
| FLExBackbone Model=Gemma-4B2026.01 | 32 | |
| FLExBackbone Model=Qwen-7B2026.01 | 39.75 | |
| FLExBackbone Model=Qwen-72B2026.01 | 51.28 | |
| FLExBackbone Model=Qwen-14B2026.01 | 82.56 | |
| FLExBackbone Model=Qwen-32B2026.01 | 83.08 |