Tencent releases Hy4 with vision and computer use; weights come later
Original title: Tencent releases Hy4 with vision and computer use, claims second place among Chinese AI models
Hy4 is API-only for now, scores 35 on Artificial Analysis, and trails Kimi K3.1 Pro by a wide margin.
Tencent released the finished Hy4 language model about two months after preview. It adds image understanding and computer use, including Blender, CAD, and game-engine tasks, but weights will not ship until a later unspecified date. Access is through Tencent Cloud, OpenRouter, and Tencent apps such as Yuanbao. Tencent’s own finance tests put Hy4 ahead of several Chinese rivals, while Artificial Analysis scores it 35, tying Grok 4.7 and second among Chinese lab models behind Kimi K3.1 Pro at 46.
Unconfirmed: Benchmark rankings and finance results are largely Tencent’s or partners’ own tests; independent Artificial Analysis scores differ.
- Pricing is unchanged: about $0.83 per million input tokens and $2.50 per million output.
- Internal web-research hallucination fell from nearly 16 percent to about 5 percent.
- Without web search, Hy4 invents an answer 89 percent of the time when it does not know.
- Hy4 with Hyra 2.0 raised internal training throughput by about 20 percent across six steps.
- On Hy-FrontierMath T5, Hy4 scored 13.0 versus 97.5 for GPT 6.1 Sol.
- Vendors
- Tencent
- Products
- Tencent CloudYuanbaoHyra 2.0
- Organizations
- Artificial AnalysisOpenRouterComeinManusTypeSafe AI
From the original article
5 min read at the-decoder.com
Tencent has unveiled Hy4, its first multimodal language model that understands images and can operate software on its own. On the Artificial Analysis Intelligence Index, it scores 35 points, solid but behind Kimi K3. 1 Pro.
By Jonathan Kemper · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.