Mathematical Reasoning on AMC 23 (Avg@16 Accuracy, Average Tool Calls (ATC))
74.7Avg@16 AccuracyQwen2.5-7B-Instruct (GRPO)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-7B-Instruct (GRPO)Backbone=Qwen2.5-7B, Training Strategy=GRPO2026.01 | 74.7 | 1.44 | |
| AdaTIR-7BBackbone=Qwen2.5-7B, Training Strategy=AdaTIR2026.01 | 72.3 | 1.02 | |
| Qwen2.5-3B-Instruct (GRPO)Backbone=Qwen2.5-3B, Training Strategy=GRPO2026.01 | 60.2 | 2.02 | |
| AdaTIR-3BBackbone=Qwen2.5-3B, Training Strategy=AdaTIR2026.01 | 58 | 1.35 | |
| Qwen2.5-7B-InstructBackbone=Qwen2.5-7B, Training Strategy=Instruct2026.01 | 52 | — | |
| Qwen2.5-3B-InstructBackbone=Qwen2.5-3B, Training Strategy=Instruct2026.01 | 38 | — | |
| Qwen2.5-3B-Instruct-TIRBackbone=Qwen2.5-3B, Training Strategy=TIR2026.01 | 30 | 1.02 | |
| Qwen2.5-7B-Instruct-TIRBackbone=Qwen2.5-7B, Training Strategy=TIR2026.01 | 21.3 | 0.51 |