DRT-o1 is a neural machine translation model that optimizes the translation process using extended reasoning chains. It extracts English sentences containing metaphors or similes and employs a multi-agent framework (including translators, advisors, and evaluators) to collaborate and construct extended reasoning machine translation samples. The DRT-o1-7B and DRT-o1-14B models are large language models trained based on Qwen2.5-7B-Instruct and Qwen2.5-14B-Instruct. The primary advantage of DRT-o1 lies in its ability to handle complex linguistic structures and deep semantic understanding, which is crucial for improving the accuracy and naturalness of machine translations.