Synthetic Time Series Reasoning on TimeSeriesExam
47.27AccuracyVeriTime
Evaluation results
| Method | Links | |
|---|---|---|
| VeriTimeBackbone=Qwen3-4B-Instruct2026.02 | 47.27 | |
| VeriTimeBackbone=Qwen2.5-3B-Instruct2026.02 | 41.67 | |
| Qwen3-4B-InstructSetting=Base2026.02 | 37.98 | |
| Qwen2.5-3B-InstructSetting=Base2026.02 | 35.66 |