List your company

Multi-hop Reasoning on MuSiQue (Accuracy)

51Accuracygemma-3-27b-it
11.27221.58631.942.214Mar 13, 2026Mar 26, 2026Apr 8, 2026Apr 21, 2026May 4, 2026May 17, 2026May 30, 2026

Evaluation results

MethodLinks
gemma-3-27b-itthinking mode=false2026.0551
Qwen3-32Bthinking mode=false2026.0549.7
Qwen3-32Bthinking mode=true2026.0549.3
Qwen3-14Bthinking mode=true2026.0545.6
gemma-3-12b-itthinking mode=false2026.0544.6
Qwen3-8Bthinking mode=true2026.0543.9
Qwen3-4Bthinking mode=true2026.0541.5
AoTBackbone=DeepSeek-V32026.0339.6
ToTBackbone=DeepSeek-V32026.0339.4
Qwen3-8Bthinking mode=false2026.0539.3
AoTBackbone=Qwen-Turbo2026.0339.2
ToTBackbone=GPT-4o-mini2026.0339.1
Qwen3-14Bthinking mode=false2026.0538.7
AoTBackbone=GPT-4o-mini2026.0338.4
ToTBackbone=Qwen-Turbo2026.0338.4
AFlowBackbone=DeepSeek-V32026.0338.2
OCC-RAG-1.7Bthinking mode=false2026.0538.2
AFlowBackbone=GPT-4o-mini2026.0338.1
TDA-RCBackbone=DeepSeek-V32026.0337.9
TDA-RCBackbone=GPT-4o-mini2026.0337.8
AFlowBackbone=Qwen-Turbo2026.0337.8
TDA-RCBackbone=Qwen-Turbo2026.0337
GoTBackbone=DeepSeek-V32026.0336.7
OCC-RAG-0.6Bthinking mode=false2026.0536.6
GoTBackbone=GPT-4o-mini2026.0336.5
GoTBackbone=Qwen-Turbo2026.0336.4
FoT (n=8)Backbone=DeepSeek-V32026.0336.2
FoT (n=8)Backbone=GPT-4o-mini2026.0335.8
FoT (n=8)Backbone=Qwen-Turbo2026.0335.8
Self-RefineBackbone=GPT-4o-mini2026.0335.1
Self-RefineBackbone=DeepSeek-V32026.0335.1
Self-RefineBackbone=Qwen-Turbo2026.0335
CoTBackbone=DeepSeek-V32026.0334.2
CoTBackbone=GPT-4o-mini2026.0334.1
CoT-SCBackbone=DeepSeek-V32026.0334
CoT-SC (n=5)Backbone=GPT-4o-mini2026.0333.8
CoTBackbone=Qwen-Turbo2026.0333.6
CoT-SC (n=5)Backbone=Qwen-Turbo2026.0333.5
Qwen3-4Bthinking mode=false2026.0533.1
Qwen3-1.7Bthinking mode=true2026.0530.7
gemma-3-4b-itthinking mode=false2026.0530.1
SmolLM3-3Bthinking mode=true2026.0529.4
SmolLM3-3Bthinking mode=false2026.0521.5
Qwen3-1.7Bthinking mode=false2026.0520.1
Qwen3-0.6Bthinking mode=true2026.0517.2
Pleias-RAG-1.2Bthinking mode=false2026.0515
Qwen3-0.6Bthinking mode=false2026.0513.2
gemma-3-1b-itthinking mode=false2026.0512.8