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
Anthropic Revenue Surges, Market Cap Nears Trillion-Dollar Mark
Reports indicate that Anthropic has achieved substantial revenue growth in just a few months. The company's annualized revenue now exceeds $30 billion—more than triple what it was at the end of last year. This growth could give Anthropic a slight edg
Apple removes Cal AI app for unauthorized in-app purchases and manipulative billing
Apple’s recent removal of Cal AI, a popular AI-powered food tracking app within MyFitnessPal, underscores its strict enforcement of App Store policies on external payments and subscriptions. The app, which generates $50 million in annual recurring re
Github Copilot's token-based billing sparks developer outrage
The golden era of Microsoft's GitHub Copilot may be ending, especially for individual users. The company is shifting from a flat subscription fee to a token-based billing model, which could significantly increase costs. While larger enterprises might
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
Anthropic Revenue Surges, Market Cap Nears Trillion-Dollar Mark
Reports indicate that Anthropic has achieved substantial revenue growth in just a few months. The company's annualized revenue now exceeds $30 billion—more than triple what it was at the end of last year. This growth could give Anthropic a slight edg
Apple removes Cal AI app for unauthorized in-app purchases and manipulative billing
Apple’s recent removal of Cal AI, a popular AI-powered food tracking app within MyFitnessPal, underscores its strict enforcement of App Store policies on external payments and subscriptions. The app, which generates $50 million in annual recurring re
Github Copilot's token-based billing sparks developer outrage
The golden era of Microsoft's GitHub Copilot may be ending, especially for individual users. The company is shifting from a flat subscription fee to a token-based billing model, which could significantly increase costs. While larger enterprises might











