# Unite.AI — Tencent Releases Full Hy4 Model With Vision and Computer Use

- Company: Unite.AI (unite.ai)
- Announced: 2026-11-04T11:18:02+00:00
- Category: new-model
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://www.unite.ai/tencent-releases-full-hy4-model-with-vision-and-computer-use/
- Record: https://forck.live/items/31502-tencent-releases-full-hy4-model-with-vision-and-computer-use
- Subject: Reporting
- Models affected: Hy4, Hy4 Flash, DeepSeek-V4.1-Flash-1017, Kimi K3.1 Pro, GLM-5.3
- Context window: 1,048,576-token context window
- Pricing: $0.834 per million input tokens, $2.501 per million output tokens and $0.042 per million cached input tokens

Tencent released the full version of Hy4 on November 4, 2026, adding vision and computer-use capabilities and serving it through hosted APIs and Tencent products, while the model weights and a system card are planned for later. The earlier Hy4 preview, open-sourced in August 2026, had 770 billion total parameters, 49 billion active parameters and a 1M-token context window, and Tencent had flagged excessive reasoning and over-verification as known issues. The full release scales pre-training and post-training, and Tencent reports gains over the preview in instruction following, long-context understanding, long-horizon tasks and deep research. Hy4 is available through Tencent Cloud and OpenRouter and inside CodeBuddy, WorkBuddy, Yuanbao and ima, at list prices unchanged from the preview. OpenRouter's listing specifies a 1,048,576-token context window and a 65,536-token completion limit, with text and image inputs and text output, making vision part of the hosted API rather than only Tencent's own applications. Tencent says the model interprets fine-grained image details and complex charts and operates across command-line and graphical interfaces, with training covering Blender, CAD, game development and finance workflows. Tencent reports gains of 19.1% on GameDev and 9.8% on GameLogic over the preview, and says Hy4 ranked first among open-weight models in a company-run finance evaluation with a score of 3.04, ahead of DeepSeek-V4.1-Flash-1017 at 2.97 and Kimi K3.1 Pro and GLM-5.3 at 2.68 each. Tencent acknowledges a gap to top proprietary models on internal benchmarks for research-level mathematics, long-horizon coding, model self-improvement and advanced visual question answering, and says a separate Hy4 Flash release targeting better cost efficiency and faster inference is planned for the near future.

## Evidence

Verbatim from https://www.unite.ai/tencent-releases-full-hy4-model-with-vision-and-computer-use/:

> Tencent announced the full release of Hy4 on November 4, 2026, adding vision and computer-use capabilities and making the model available through hosted APIs and Tencent products. The model weights and a comprehensive system card are planned for a later release.
> 
> The earlier Hy4 preview , by contrast, was open-sourced when it arrived in August 2026. It had 770 billion total parameters, 49 billion active parameters and a 1M-token context window, and Tencent identified excessive reasoning and over-verification as known issues in that early version.
> 
> The full release substantially scales both pre-training and post-training, with Tencent reporting gains over the preview in instruction following, long-context understanding, long-horizon tasks and deep research.
> 
> Hosted Access at Unchanged List Prices
> 
> Hy4 is available through Tencent Cloud and OpenRouter and inside CodeBuddy, WorkBuddy, Yuanbao and ima. Listed API pricing is unchanged from the preview at $0.834 per million input tokens, $2.501 per million output tokens and $0.042 per million cached input tokens.
> 
> OpenRouter’s deployment listing specifies a 1,048,576-token context window and a 65,536-token completion limit.

---

Record: https://forck.live/items/31502-tencent-releases-full-hy4-model-with-vision-and-computer-use
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md