List your company
ResearchFormerly WizWand

VRP optimization-modeling on VRP Optimization-Modeling Benchmark 1

96.13Pass@1VRPCoder-GRPO
-3.845222.109948.06574.0201Jun 3, 2026
Updated 4mo ago

Evaluation results

MethodLinks
VRPCoder-GRPOModel scale=8B, Stage=GRPO2026.0696.13
Gemini-3.1-Pro PreviewModel type=Closed-source frontier LLM2026.0695.81
VRPCoder-SFTModel scale=8B, Stage=SFT2026.0692.26
Claude-Sonnet-4.5Model type=Closed-source frontier LLM2026.0670.65
DeepSeek-V3.2Model type=Open-source general-purpose LLM2026.0631.29
MiniMax-M2.5Model type=Open-source general-purpose LLM2026.0621.61
Qwen2.5-72B-InstructModel type=Open-source general-purpose LLM2026.0615.81
SIRL-Gurobi-32BModel type=OR-LLM2026.068.71
ORLM-LLaMA-3-8BModel type=OR-LLM2026.063.87
SIRL-Gurobi-8BModel type=OR-LLM2026.061.94
OptMATH-Qwen2.5-7BModel type=OR-LLM2026.060.97
Qwen3-8BModel type=Open-source general-purpose LLM2026.060