Model Releases

Tencent Hunyuan compresses on-device translation model Hy-MT2-1.8B to 440MB, deployed in Bilibili live danmaku translation

Original title:3.3GB 压到 440MB,混元极低 bit 翻译模型跑进直播弹幕

AI summary

Tencent Hunyuan compressed its on-device translation model Hy-MT2-1.8B to 574MB and 440MB using 2-bit and 1.25-bit quantization schemes, with almost no loss in translation quality and outperforming commercial APIs such as Microsoft Translator on FLORES-200. The model has completed x86 adaptation with Intel and is deployed in real-time Bilibili live danmaku translation, taking 500–800ms per danmaku.

Key Takeaways

  • Tencent Hunyuan compressed its on-device translation model Hy-MT2-1.
  • 8B to 574MB and 440MB using 2-bit and 1.
  • 25-bit quantization schemes,with almost no loss in translation quality and outperforming…
  • The model has completed x86 adaptation with Intel and is deployed in real-time Bilibili live…

Quick Facts

Published
Updated
Source
公众号:腾讯混元
Category
Model Releases
AI score
67
Read the original Read the full article at the original source
Why it matters

相比精度数字,这套量化给出 x86 优化和直播弹幕实测数据,让端侧替代云翻译 API 的成本与隐私权衡有了可对照的工程基线。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Tencent Hunyuan compresses on-device translation model Hy-MT2-1.8B to 440MB, deployed in Bilibili live danmaku translation. August 26, 2026. https://new.nat6.net/en/news/cmta7honk0553roj21gque0zw/

View all →