Zhipu AI has launched the MathGLM mathematical model, aimed at enhancing the performance of large models in mathematical reasoning. MathGLM can accurately compute arithmetic operations, including basic arithmetic and complex mixed operations. MathGLM has shown excellent performance in solving math problems in Chinese. Its performance on arithmetic tasks surpasses that of GPT-4. MathGLM has achieved good results in addressing math problems for different grade levels and solving practical math problems.

Researchers from Tsinghua University and others have introduced MathGLM, a 2 billion parameter language model that outperforms GPT-4 with nearly 100% accuracy in arithmetic tasks. MathGLM utilizes a decoder-only architecture and has significantly improved its mathematical capabilities through training on a large-scale arithmetic dataset. It also surpasses models like GPT-4 and ChatGPT in handling complex mixed arithmetic operations with intricate numerical formats.