Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
githu.com-2111 presents the MuMath-Code project, a series of tool-use large language models designed to enhance mathematical reasoning. These models are trained using a two-stage strategy combining multi-perspective data augmentation with code generation and execution, achieving improved performance on reasoning benchmarks compared to existing open-source and proprietary methods.
Parse Score