Tencent quietly released the Hunyuan Hy3 preview: the first flagship model since Yao Shunyu joined.

ME News message: On April 23 (UTC+8), according to Dongcha Beating monitoring, Tencent released the latest flagship model in its Hunyuan series, Hy3 preview. It adopts an MoE (Mixture of Experts—multiple small expert networks activated on demand) architecture, with a total of 295 billion parameters and 21 billion activated parameters. The model has been made available via the OpenRouter API and a HuggingFace demo, but open-source weights have not yet been released. This is the first flagship model led by Yao Shunyu, a former OpenAI researcher and Tencent’s Chief AI Scientist, after he joined—and also the result of Tencent’s work in producing outputs in under 3 months after it rebuilt its basic model research and development system in February this year. Tencent itself positions it as “comparable to top domestic models, but still behind cutting-edge American models such as OpenAI and Google DeepMind.” The model is primarily focused on agent capabilities, with dedicated adaptation and optimization for mainstream agent tools such as OpenClaw. It supports a 262.1K context window and a three-level reasoning mode. It has already been deployed to Tencent products such as Yuanbao and CodeBuddy. (Source: BlockBeats)
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned