Tencent releases Hy4: first vision and computer-use support, aimed at the top tier of open models
According to the South China Morning Post, Tencent released its Hy4 large language model on November 4 with visual understanding for the first time: it processes images, analyses complex charts and does visual mathematical reasoning. Tencent's announcement says Hy4 scales up both pre-training and post-training compared with the Hy4 preview, supports computer use through CLI and GUI, and is available on Tencent Cloud, OpenRouter and in CodeBuddy, WorkBuddy, Yuanbao and ima at an unchanged API price. Against Hy4 preview, Tencent reports GameDev up 19.1%, GameLogic up 9.8%, and the hallucination rate on search tasks down from 15.76% to 5.45% in internal evaluations. Tencent acknowledges Hy4 still trails the leading closed systems on research-level maths, long-horizon coding, model self-improvement and advanced visual question answering.
Source: South China Morning Post
Need an LLM chatbot or agent for your business?
Describe your need briefly; you will get clarifying questions and a scoped proposal before any work starts.
Related news
LeanGuard: checking every tool call of an LLM agent against a policy written in Lean 4
Strategist: a meta-agent that evolves inference strategies as code, with 39% lower inference cost
