I'm LongbridgeAI. Here is the key information I extractedZhitong Finance APP learned that on November 4, Tencent Hunyuan released the official version of Hy4. With another comprehensive leap in model capabilities, Hy has joined the ranks of the most powerful open-source models for the first time. Compared to Hy4 preview, Hy4 has significantly expanded the scale of pre-training and post-training, introduced visual capabilities for the first time, and substantially improved various comprehensive abilities including instruction following, long-text understanding, long-horizon task execution, and deep research.
Hy4 introduces visual capabilities for the first time, combining perception, reasoning, and tool invocation to perform tasks such as granular image processing, complex chart analysis, and mathematical visual reasoning. Hy4 also supports computer operations for the first time, allowing flexible combination of command-line interfaces and graphical user interfaces. Tencent Hunyuan has conducted specialized training for professional software such as Blender and CAD, as well as cross-application long-horizon tasks, continuously optimizing planning and execution capabilities for professional tasks.
The introduction of visual understanding has also significantly enhanced Hy4's front-end and game development capabilities. Tencent Hunyuan has performed targeted optimizations for complex interactions, business logic, state management, and front-end/back-end integration in front-end development, ensuring that the interfaces generated by the model are not only visually appealing but also fully functional and usable. In game development scenarios, Tencent Hunyuan collaborates with professional game developers from Tencent IEG to continuously optimize data and evaluation quality, focusing on improving capabilities in game feature implementation, gameplay logic, state understanding, and asset processing. This has achieved deep adaptation to mainstream engines such as UE, Unity, and Godot, and has begun expanding into multiplayer online game development. Compared to Hy4 preview, Hy4 showed improvements of 19.1% and 9.8% in GameDev and GameLogic evaluations, respectively.
Tencent Hunyuan has constructed thousands of workspaces and tens of thousands of high-quality tasks based on real financial scenarios, enabling the model to thoroughly learn professional processes and standards in areas such as fund research, securities analysis, quantitative strategies, asset allocation, corporate finance, and fiscal and tax accounting. Tencent Hunyuan, in collaboration with Tencent Financial Technology (FiT), conducted financial crowd-testing with 213 experts, 153 of whom were from leading external financial institutions, with 70% having more than five years of work experience. Results based on 579 real financial tasks show that Hy4 ranks first among open-source models in overall performance (crowd-testing scores: Hy4 3.04, Kimi K3.1 Pro 2.68, GLM-5.3 2.68, DeepSeek-V4.1-Flash-1017 2.97). Furthermore, in the internal IRAB evaluation by Xuntu Technology, a financial AI agent company, Hy4 improved by 12.03 percentage points compared to Hy4 preview.
To enhance practicality in financial and other office scenarios, Tencent Hunyuan has initiated specialized search optimizations targeting timely retrieval, multi-source information integration, and factual consistency verification. In internal testing, Hy4's hallucination rate decreased from 15.76% in the preview version to 5.45%, and the error rate caused by guessing dropped from 23.1% to 5.6%. Additionally, Hy4 has significantly improved human-AI co-writing capabilities in office scenarios, supporting real-time collaborative editing of Office documents such as Word, Excel, and PPT with users. Blind evaluations show that Hy4 has made significant improvements over Hy4 preview in document aesthetics, stability, and usability.
By combining complex reasoning, scientific programming, and long-horizon task execution, Hy4 further enhances its ability to drive scientific research end-to-end. In the internal TenureTrail evaluation, which covers 7 major disciplines and 123 executable research tasks, Hy4 demonstrated solid scientific research capabilities in fields such as mathematical construction, physical modeling, earth and environment, systems engineering, and socio-economic decision-making, ranking first among open-source models.
Hy4 is also preparing for Recursive Self-Improvement (RSI) to enhance capabilities in various model development tasks. Combined with the Hyra2.0 scaffolding, Hy4 optimized the pre-training infrastructure for smaller models with the same internal architecture and automatically trained the Jev model, which outperforms the official version in multiple evaluations. Meanwhile, Tencent Hunyuan has open-sourced HyCell, a single-cell representation foundation model, and HyPDE, a physical field foundation model, both developed by Hy4 in collaboration with Hyra2.0, aiming to translate model-assisted scientific research outcomes into reusable tools for the community.