Home
Tencent Hunyuan Launches 0.4G Offline Translation Model, Enabling Signal-Free Communication for May Day Travel
With the May Day Golden Week approaching, the Tencent Huan Yuan team has released an ultra-quantized edition of their professional translation large model, offering a solid language solution for travelers heading overseas. Named Hy-MT1.5-1.8B-1.25bit, the model is just 440MB in size and runs efficiently on mobile devices locally.

Ultra-Compression Technology Enables Offline Use on Mobile
The model uses advanced quantization to compress its original size from over 3GB to just one-eighth, allowing it to operate smoothly in the background on standard smartphones. Even in remote scenic locations with no network signal or during sudden emergencies, users can still get instant, high-quality translations.
The team offers two quantization options: 2-bit and 1.25-bit. The 1.25-bit version employs a fine-grained sparsity strategy accepted at a top academic conference, where three out of every four parameters are stored using 1 bit, ensuring strong translation performance despite the compact size.

Professional-Grade Translation Quality with Privacy Protection
Despite its small footprint, the model delivers outstanding performance across 33 languages, five dialects, and thousands of translation directions. On multiple authoritative benchmarks, its translation accuracy exceeds that of Google Translate and rivals commercial models with far more parameters.
Because all translation processing happens entirely on the device, users don't have to worry about personal sensitive data being collected or uploaded — significantly enhancing privacy and security. The project is now fully open source and includes a demo app, allowing Android users to try features like background word lookup with no subscription fees.
Try the demo:
Hugging Face (for users abroad): https://huggingface.co/AngelSlim/Hy-MT1.5-1.8B-1.25bit-GGUF/resolve/main/Hy-MT-demo.apk
ModelScope Community (for users in China): https://modelscope.cn/models/AngelSlim/Hy-MT1.5-1.8B-1.25bit-GGUF/resolve/master/Hy-MT-demo.apk
Related article
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage
California AV Compliance: A New Era of Tickets, Geofences, and 1M Miles
Guident operates an AuveTech shuttle in South Florida, managing a four-mile route in West Palm Beach and a one-mile route in Boca Raton using its remote monitoring technology. | Credit: GuidentCalifornia is redefining the regulatory landscape for dri
Related Special Topic Recommendations
Comments (0)
0/500
With the May Day Golden Week approaching, the Tencent Huan Yuan team has released an ultra-quantized edition of their professional translation large model, offering a solid language solution for travelers heading overseas. Named Hy-MT1.5-1.8B-1.25bit, the model is just 440MB in size and runs efficiently on mobile devices locally.

Ultra-Compression Technology Enables Offline Use on Mobile
The model uses advanced quantization to compress its original size from over 3GB to just one-eighth, allowing it to operate smoothly in the background on standard smartphones. Even in remote scenic locations with no network signal or during sudden emergencies, users can still get instant, high-quality translations.
The team offers two quantization options: 2-bit and 1.25-bit. The 1.25-bit version employs a fine-grained sparsity strategy accepted at a top academic conference, where three out of every four parameters are stored using 1 bit, ensuring strong translation performance despite the compact size.

Professional-Grade Translation Quality with Privacy Protection
Despite its small footprint, the model delivers outstanding performance across 33 languages, five dialects, and thousands of translation directions. On multiple authoritative benchmarks, its translation accuracy exceeds that of Google Translate and rivals commercial models with far more parameters.
Because all translation processing happens entirely on the device, users don't have to worry about personal sensitive data being collected or uploaded — significantly enhancing privacy and security. The project is now fully open source and includes a demo app, allowing Android users to try features like background word lookup with no subscription fees.
Try the demo:
Hugging Face (for users abroad): https://huggingface.co/AngelSlim/Hy-MT1.5-1.8B-1.25bit-GGUF/resolve/main/Hy-MT-demo.apk
ModelScope Community (for users in China): https://modelscope.cn/models/AngelSlim/Hy-MT1.5-1.8B-1.25bit-GGUF/resolve/master/Hy-MT-demo.apk
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage











