Tencent releases Hy4: first vision understanding and computer use, unchanged pricing, weights promised within days
Tencent formally released its next-generation Hy4 model on 4 November, adding image understanding and computer use for the first time at the same price as August’s Hy4 Preview; the company says weights and system cards will follow within days. On independent testing, Artificial Analysis scores it 35, second among Chinese labs’ models.
Tencent says Hy4 handles fine-grained images, complex charts and mathematical visual reasoning, and can operate computers through command lines and GUIs for the first time; training specifically covered professional software such as Blender and CAD and tasks spanning several applications. It is available on Tencent Cloud TokenHub, WorkBuddy, CodeBuddy, Yuanbao and WeCom’s AI assistant, plus API access through third parties such as OpenRouter.
Pricing matches Hy4 Preview: about $0.90 per million input tokens, $2.70 output and $0.045 cached input (The Decoder converts this to roughly $0.83 and $2.50). WorkBuddy and CodeBuddy are free from 4 to 15 November, and Hy4 on TokenHub is 50% off through 31 December; a cheaper Hy4 Flash is coming. Tencent says weights and system cards will be released within days, with no date given.
In Tencent’s own test of 579 real finance tasks, Hy4 scored highest overall (3.04 versus 2.97 for DeepSeek-V4.1-Flash-1017 and 2.68 each for Kimi K3.1 Pro and GLM-5.3). Independent results are more cautious: Artificial Analysis gives it 35, above the median of 30 but behind Kimi K3.1 Pro’s 46, and on a no-search factual test it invents an answer 89% of the time when it does not know. Tencent also admits it still trails the best commercial models on hard math, long coding projects and complex image analysis.
Sources
- Tencent Launches Hy4, Joining the Ranks of the World’s Leading Open-Source AI Models — Tencent
- Tencent releases Hy4 with vision and computer use, claims second place among Chinese AI models — The Decoder