Qwen-MT更新：支持92种语言，低成本高速度翻译
原标题：Qwen-MT: Where Speed Meets Smart Translation
AI 摘要
Qwen团队通过Qwen API发布了Qwen-MT（qwen-mt-turbo）的最新更新，该模型基于Qwen3，支持92种语言，覆盖全球95%以上人口。新版本通过强化学习提升了翻译准确性和流畅度，并提供了术语干预、领域提示和翻译记忆等高级定制功能。Qwen-MT采用轻量级MoE架构，API成本低至每百万输出token 0.5美元，在自动评估和人工评估中均优于GPT-4.1-mini、Gemini-2.5-Flash等同类模型，并与GPT-4.1、Gemini-2.5-Pro等大型模型保持竞争力。
正文节选
Introduction Here we introduce the latest update of Qwen-MT (qwen-mt-turbo) via Qwen API. This update builds upon the powerful Qwen3, leveraging trillions multilingual and translation tokens to comprehensively enhance the model’s multilingual understanding and translation capabilities. By integrating reinforcement learning techniques, the model achieves significant improvements in translation accuracy and linguistic fluency. Key Features: - Multilingual Support for 92 Languages: Qwen-MT enables