Multi-Turn Tool-Integrated Reasoning (TIR) on MATH500
84.69Peak avg@32 ScoreOTB
Evaluation Results
| Method | Links | |
|---|---|---|
| OTBBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 84.69 | |
| SimpleTIRBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 82.25 | |
| OPOBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 76.94 | |
| GRPOBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 76.88 | |
| RLOOBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 76.06 | |
| OGBBackbone=Qwen2.5-7B, Reasoning Paradigm=TIR2026.02 | 75.69 |