Insights: quantization

Technical insights tagged with "quantization".

Filter by Tags

All 2025 2026 3D Generation A2UI ACP AI AI Agents AI Audio AI Image AI Image Generation AI Models AI Tools AI Video AI Video Creation AI Vision AP2 ASR Agents Alibaba Automation Baidu Benchmark Bilibili ByteDance CBT CLI CLI Tools Chinese NLP Chrome Claude Coding Coding Agent Commerce Computer Use Configuration Content Creation CosyVoice Creative DeepSeek Design Document Understanding GLM GLM-5 GLM-OCR GUI Agent Gemini Gemini 3 Google Google AI Graphviz Hunyuan Image Editing Image Generation Image Model January 2026 Kimi LLM Language Models LoRA Local LLM Machine Learning Math Memory Mental Health Microsoft MoE Moltbook Moonshot Multimodal NVIDIA Nano Banana Pro New Websites OCR Open Source OpenClaw Payments Personality Photo Photo Editing Platform Premium Apps Productivity Prompts Protocols Qwen Reasoning Seedance Social Network Sora TTS Tencent Tools Translation Trending Tools Tutorial Tutorials UCP Vibe Coding Video Video Creation Video Generation Vision Vision-Language Model WebMCP Weekly Picks Xiaomi Z-Image Turbo Zhipu AI color design game machine-learning multilingual on-device open-source perception quantization translation ux

April 30, 2026

Hy-MT1.5-1.8B-2bit: Tencent Open-Sources a 574MB On-Device Translation Model That Beats 72B Giants

Tencent Hunyuan Team releases Hy-MT1.5-1.8B-2bit, a breakthrough 2-bit quantized translation model. At just 574MB, it outperforms Tower-Plus-72B and commercial APIs across 33 languages. Full offline on-device deployment available.

AItranslationTencent+4