Tencent's New Translation Models Outperform Google Translate, GPT-4.1
Tencent has launched two groundbreaking translation models, Hunyuan-MT-7B and Hunyuan-MT-Chimera-7B, making them open source. These models have set new benchmarks in multilingual translation, outperforming established systems like Google Translate and GPT-4.1.
Tencent's models support bidirectional translation between 33 languages, including widely spoken and less frequently digitized ones. They excel in translation between Mandarin Chinese and ethnic minority languages in China, such as Kazakh, Uyghur, Mongolian, and Tibetan. The company's five-stage training process combines several AI methods to achieve these results.
At the recent WMT2025, Tencent's models achieved the best results in 30 out of 31 tested language combinations. Despite having only 7 billion parameters, they match or surpass the performance of larger foundational models. The Chimera model uses a fusion approach to combine multiple translation suggestions, improving the final output.
Tencent's Hunyuan-MT-7B and Hunyuan-MT-Chimera-7B models are now available as open source on Hugging Face and GitHub, offering a significant advancement in multilingual translation. Their performance in international tests and focus on less digitized languages make them valuable tools for global communication.
Read also:
- Mural at blast site in CDMX commemorates Alicia Matías, sacrificing life for granddaughter's safety
- Autonomous Vehicles Hit Manhattan & Brooklyn Streets, Sparking Controversy
- Increased energy demand counters Trump's pro-fossil fuel strategies, according to APG's infrastructure team.
- AI-Powered Transportation Stock's Possible Challenge to Tesla's Autonomous Dreams?
