Translated data: The 7B open-source mathematical model developed by the Chinese team has outperformed the trillion-scale GPT-4, achieving a competition-level accuracy of 51.7% without the aid of external tools. This model has broken through the limits of open-source models and has received high recognition within the industry. By utilizing advanced technologies and large-scale datasets, the model has achieved stunning results in mathematical capabilities, demonstrating immense potential and superiority. The entire open-source community is amazed by its performance, prompting a reevaluation of the strength of open-source mathematical models.