qwen-mt-plus

Qwen-MT is an upgraded translation model developed by Alibaba Tongyi based on Qwen3. Its core positioning is as a “high-speed, low-cost, multilingual intelligent translation assistant,” achieving high-quality bidirectional translation through a lightweight architecture and catering to global language communication needs across various scenarios.

• Ultra-wide language coverage: Supports bidirectional translation among 92 languages, covering major official languages and important dialects, reaching over 95% of the global population and adapting to diverse multilingual communication scenarios.
• High-performance translation results: Outperforms models of similar size in Chinese-English, English-German, and other domains, as well as in the WMT24 task. It leads in both pass rate and excellent rate in human evaluations, producing translations that are accurate and fluent.
• Highly controllable and flexible: Offers features such as terminology intervention, domain-specific prompts, and memory libraries, allowing users to customize translation rules and tailor the model to professional and specific application needs.
• Efficient and cost-effective: Employs a lightweight MoE architecture with low response latency; API calls cost as little as 2 yuan per million output tokens, making it ideal for high-concurrency and real-time applications.

───────────────────────────────────────────────────────────────────

Core Capabilities

🌍 Accurate multilingual bidirectional translation: Covers a wide range of languages, ensuring “faithfulness, fluency, and elegance” whether in everyday conversations, cultural content, or specialized texts.

🎯 Adaptation to professional scenarios: Supports customized translation in fields such as law, IT, and government affairs, with strong consistency in terminology and alignment with industry-specific expression conventions.

⚡ Real-time and efficient response: The lightweight architecture ensures low latency, meeting the demands of real-time communication and high-concurrency translation, offering outstanding cost-effectiveness.

✏️ Flexible and customizable configuration: Allows intervention in terminology and setting translation styles, adapting to personalized and complex scenarios and enhancing translation precision.

🔌 Easy integration and deployment: Compatible with OpenAI SDK standards, providing concise call codes for quick integration into various applications and workflows.

Note: This model requires specifying the target languages to be translated via the"translation_options" parameter. For details on available languages, please refer to the documentation.