Hybrid Reasoning on CLUTRR (test)
76.4AccuracyAgent-GWO
Evaluation results
| Method | Links | |
|---|---|---|
| Agent-GWOBackbone=GPT-4o-mini2026.04 | 76.4 | |
| AoTBackbone=GPT-4o-mini2026.04 | 75.2 | |
| GoTBackbone=GPT-4o-mini2026.04 | 74.2 | |
| ToTBackbone=GPT-4o-mini2026.04 | 73.6 | |
| AFlowBackbone=GPT-4o-mini2026.04 | 72.8 | |
| CoT-SC/n=5Backbone=GPT-4o-mini2026.04 | 72.2 | |
| Self-RefineBackbone=GPT-4o-mini2026.04 | 68.4 | |
| CoTBackbone=GPT-4o-mini2026.04 | 66.2 | |
| Agent-GWOBackbone=Gemma-3-12b-it2026.04 | 56.9 | |
| AoTBackbone=Gemma-3-12b-it2026.04 | 54.8 | |
| GoTBackbone=Gemma-3-12b-it2026.04 | 53.7 | |
| AFlowBackbone=Gemma-3-12b-it2026.04 | 53 | |
| ToTBackbone=Gemma-3-12b-it2026.04 | 52.6 | |
| CoT-SC/n=5Backbone=Gemma-3-12b-it2026.04 | 52 | |
| Self-RefineBackbone=Gemma-3-12b-it2026.04 | 50.6 | |
| CoTBackbone=Gemma-3-12b-it2026.04 | 49.3 | |
| Agent-GWOBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 27.8 | |
| AoTBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 26.4 | |
| GoTBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 23.8 | |
| ToTBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 23 | |
| AFlowBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 22.7 | |
| CoT-SC/n=5Backbone=Qwen2.5-Coder-7B-Instruct2026.04 | 20.8 | |
| Self-RefineBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 20.8 | |
| CoTBackbone=Qwen2.5-Coder-7B-Instruct2026.04 | 20.4 |