腾讯混元 Hunyuan-MT
Open-source translation model — won 30/31 language directions at WMT25
An open-source translation large model that took first place in 30 of 31 language directions at the WMT25 international translation competition.
Translate documents, parse PDFs, and extract text with OCR.
Key features
- Translation model Hunyuan-MT-7B plus ensemble model Hunyuan-MT-Chimera
- Supports mutual translation among 33 languages incl. 5 Chinese minority languages
- Full training framework: pretrain, CPT, SFT, translation RL, ensemble RL
- FP8 quantized variants available
Good for
- Developers needing open-source machine translation at 7B scale
- Teams translating Chinese minority languages
Stack & tags
PythonTransformersmachine-translationmultilingualopen-weightsensemble7btransformers
Quick start
pip install transformers==4.56.0From the project README on GitHub.
Similar open-source projects
微软 MarkItDown
★ 187,737Microsoft's open-source Python tool — convert PDF/Word/PPT/audio/YouTube to clean Markdown
DeepSeek-OCR
★ 23,926DeepSeek's novel OCR document model — "contextual optical compression"