Tencent Launches Hy4 Open-Source AI Model
Original title: Tencent Launches Hy4, One of the Leading Open-Source AI Models
Hy4 adds visual understanding and computer-use capabilities and will release weights soon.
Tencent announced the release of Hy4, its latest large language model, which ranks among the world’s leading open-source AI models. The model features major improvements in instruction following, long-context comprehension, and long-horizon agentic tasks, including first-time visual understanding for fine-grained image processing and chart analysis. Hy4 also introduces computer-use capabilities to interact with command-line and graphical interfaces, and has been trained on professional software like Blender and CAD.
- Hy4 supports fine-grained image processing, complex chart analysis, and mathematical visual reasoning
- Hy4 enables interaction with CLI and GUI for work and everyday tasks
- Hy4 optimized for financial analysis with top scores in blind evaluation by 213 professionals
- Hy4 improved GameDev by 19.1% and GameLogic by 9.8% in benchmarks with Unreal Engine, Unity, and Godot
- Hy4 reduces hallucination rate to 7.2% and unsupported assumption errors to 5.6%
From the original article
5 min read at tencent.com
Tencent today announced the release of Tencent Hy4, its latest-generation large language model, marking a significant leap in performance and capabilities that now ranks it among the world’s leading open-source AI models.
By gmays · tencent.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.