Reasoning on 10 challenging reasoning tasks Combined
46.7Average ScoreE3-TIR
Evaluation Results
| Method | Links | |
|---|---|---|
| E3-TIRBackbone=Qwen2.5-3B, Agent Category=Multi-Tool Integrated Agent2026.04 | 46.7 | |
| ARPOBackbone=Qwen2.5-3B, Agent Category=Multi-Tool Integrated Agent2026.04 | 45.2 | |
| Tool-StarBackbone=Qwen2.5-3B, Agent Category=Multi-Tool Integrated Agent2026.04 | 44.3 | |
| ReCallBackbone=Qwen2.5-3B, Agent Category=Multi-Tool Integrated Agent2026.04 | 40.5 | |
| Tree-GRPOBackbone=Qwen2.5-3B, Agent Category=LLM-based Search Agent2026.04 | 36.1 | |
| Search-R1Backbone=Qwen2.5-3B, Agent Category=LLM-based Search Agent2026.04 | 35.6 | |
| ToRLBackbone=Qwen2.5-3B, Agent Category=LLM-based Math Agent2026.04 | 32.2 | |
| SimpleTIRBackbone=Qwen2.5-3B, Agent Category=LLM-based Math Agent2026.04 | 31.7 | |
| Search-o1Backbone=Qwen2.5-3B, Agent Category=LLM-based Search Agent2026.04 | 30.7 | |
| Qwen2.5-3B-InstructBackbone=Qwen2.5-3B2026.04 | 26.8 |