China’s Tencent releases new open-source AI model for coding, research tasks

Aug 28 (Reuters) – Tencent (0700.HK), opens new tab released a preview version of a new open-source AI model aimed at tasks such as software ‌engineering, research and financial analysis, according to its post published on Friday on Hugging Face, a repository of open-source AI models.
 
  • The model, called Hy4 preview, uses ⁠a “mixture-of-experts” design with 770 billion parameters ​in total, Tencent said, although only about ​49 billion are used for any given text request.
  • The Chinese technology giant said it plans ​to integrate the model alongside Tencent ​products, including CodeBuddy and WorkBuddy.
  • Tencent added that the ‌early-release ⁠model can sometimes take longer than necessary to work through complex questions and may over-verify its own answers.
  • The release ​comes at ​a time ⁠when Tencent is ramping up investment in AI in ​an increasingly crowded market.
  • It unveiled a ​large ⁠language model called Hunyuan 3.0 in April, its first major release since hiring ⁠former ​OpenAI researcher Yao Shunyu ​to lead its AI platform development.