Tencent Hunyuan has officially released its Hy4 model, following a preview version that was available since late August 2026. The author details the necessary changes for an OpenAI-compatible client to interact with Hy4, focusing on the `model` field and potential issues with empty responses. The new formal version introduces vision capabilities and claims of computer use, allowing image and text to be processed within the same message structure. AI
IMPACT This release introduces multimodal capabilities and potential improvements in agentic reasoning, requiring developers to update their clients for seamless integration.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →