HONG KONG — Tencent has expanded international availability of its latest artificial intelligence model, Hy3, making the technology accessible to developers, enterprises, and individual users through a growing range of cloud services and productivity platforms.
The expansion follows the model’s official debut in July and represents Tencent’s broader strategy to accelerate enterprise AI adoption by integrating advanced reasoning and autonomous agent capabilities into business workflows. Global users can currently access Hy3 through WorkBuddy, Tencent Design Miora, and Tencent Cloud TokenHub, while additional integrations across Tencent Cloud’s AI ecosystem are expected in the coming months.
As part of the rollout, Tencent is offering free access to Hy3 through WorkBuddy for users worldwide until 31 August 2026, allowing organizations and developers to evaluate the platform before wider commercial deployment.
According to Tencent Cloud International Senior Vice President Poshu Yeung, the company aims to move businesses beyond AI experimentation by providing scalable, secure, and enterprise-ready intelligence that can be deployed across real-world operations.
Tencent reported strong early adoption following the model’s release, with Hy3 generating more than 68 times the API traffic of its predecessor. Within a week of launch, the model also reached the top position on OpenRouter’s global large language model usage rankings. The company has continued expanding Hy3 across third-party developer ecosystems, including Hermes, Kilo, Cline, OpenClaw, OpenCode, and Cherry Studio, while also releasing it on open-source platforms such as Hugging Face and ModelScope.
Designed for Enterprise-Scale AI
Hy3 is powered by a hybrid fast-and-slow-thinking Mixture-of-Experts (MoE) architecture featuring 295 billion total parameters, with 21 billion active parameters during inference. The model supports context windows of up to 256,000 tokens, enabling it to process large documents and complex reasoning tasks.
Tencent says the model delivers performance comparable to much larger AI systems across reasoning, coding, instruction following, contextual learning, and autonomous agent operations while using fewer computational resources.
Unlike models developed solely for research, Hy3 has been trained and refined using real-world feedback collected from Tencent’s own products and millions of active users. This continuous deployment cycle has helped improve both reliability and performance across demanding enterprise environments.
The company says Hy3 has demonstrated strong results in software engineering, workplace productivity, financial analysis, frontend development, and game production while lowering token consumption for many tasks.
AI Integrated Into Workplace Productivity
Tencent is positioning WorkBuddy as one of the primary interfaces for Hy3. The AI-powered workspace enables users to complete research, analyze data, prepare reports, generate presentations, and automate workplace communication through natural language prompts.
Its multi-agent architecture supports more than 100 built-in skills, allowing multiple AI agents to collaborate on complex assignments from a single request. Developers can also integrate external models through APIs and manage workflows via platforms such as Discord, Slack, and Telegram.
Internal testing conducted by Tencent showed that Hy3 achieved a task completion success rate exceeding 90% within WorkBuddy while reducing average execution time by approximately 34% compared with the previous Hy model.
Creative and Cloud Applications
Tencent has also integrated Hy3 into Tencent Design Miora, an AI-native creative platform designed for marketing teams, designers, and digital content creators.
The platform combines specialized AI agents capable of transforming simple text prompts into production-ready creative assets, including graphics, videos, 3D content, and user interface designs. Tencent says Hy3’s enhanced reasoning capabilities improve creative workflows while maintaining transparency and editability throughout the design process.
For enterprise customers, Hy3 is available through Tencent Cloud TokenHub, the company’s Model-as-a-Service platform that enables organizations to deploy and manage multiple large language models using a unified API. The platform automatically routes workloads to the most suitable model while offering centralized governance, flexible pricing, and token management tools.
Tencent is also expanding Hy3 through regional technology partnerships, including collaborations with South Korea’s e-commerce platform provider Cafe24 and Japan-based enterprise AI company Metelix, allowing businesses in those markets to access the model through existing AI gateway services.
Open-Source Strategy and Global Expansion
Reinforcing its commitment to an open AI ecosystem, Tencent has released Hy3 under the commercially permissive Apache 2.0 license, enabling developers to modify, deploy, and integrate the model into their own applications.
The company has also introduced competitive API pricing through OpenRouter, positioning Hy3 as a cost-effective option for organizations building AI-powered applications and autonomous workflows.
Supporting the global rollout is Tencent Cloud’s international infrastructure, which spans 66 availability zones across 23 geographic regions and is backed by more than 3,200 global acceleration nodes. Tencent says this infrastructure allows enterprises across Asia-Pacific, Europe, the Middle East, and the Americas to deploy Hy3 while meeting regional performance and operational requirements.
With broader global availability, Tencent is strengthening its position in the increasingly competitive enterprise AI market, betting that practical deployment, open access, and integrated cloud services will drive the next wave of artificial intelligence adoption.
