Tencent's Hy4 multimodal model with visual understanding, computer use, 3D/Blender and knowledge-work abilities.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Hy4 is a model from Tencent's Hy line, presented on Tencent's research site. According to a YouTube creator's summary, it combines visual understanding, computer use, 3D/Blender tasks and knowledge-work capabilities, and Tencent positions it among the stronger open models. These claims come from a secondary source and have not been verified against the official page text, which was not retrievable in detail.
Access is reported to be through Tencent Cloud and Tencent's WorkBuddy product. Weights were said to be due "in coming days" at the time of the video, so open-weights availability and licence terms should be confirmed on release. Parameter count, context length, benchmarks and pricing are not stated in the available material.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Tencent's new model with visual understanding, computer use, 3D/Blender and knowledge-work abilities, claimed among the best open models. Weights due in coming days; available via Tencent Cloud and WorkBuddy.
Tencent's best open-source model: 770B MoE with 49B active under Apache 2, using gated DeepSeek sparse attention. It beats Qwen 3.8 Max and DeepSeek v4 Pro and ties GLM 5.3 Flash on Arena web dev.
Closest neighbours by meaning, not just by category.
Model · Tencent
Tencent's open-weights 295B Mixture-of-Experts model (21B active) for reasoning and agentic coding.
Model · Tencent
Tencent's unified multimodal model for 3D understanding, text-to-3D generation, instruction-based editing and part generation.
Model · Tencent
Tencent Hunyuan interactive world model generating real-time, controllable 720p video with memory for consistency.
Model · Tencent
Tencent's open multilingual translation model family (1.8B, 7B, 30B-A3B MoE) covering 33 languages.
Model · Mistral
Mistral's roughly 1-trillion-parameter multimodal mixture-of-experts model with open weights, trained in Europe.
Model · Tencent
Tencent Hunyuan's real-time Chinese-English speech-to-speech simultaneous translation model with voice cloning.
