“Over the past year, we have dedicated significant efforts to researching and enhancing the reasoning capabilities of large language models, with a particular focus on their ability to solve ...