Tencent to expand AI model lineup with Hy4 and Hy4 Flash
Tencent announced the upcoming release of its Hy4 and Hy4 Flash AI models, which will incorporate visual processing capabilities into its existing architecture.
Tencent is preparing to launch its Hy4 and Hy4 Flash AI models, according to an announcement made by senior executive vice president Dowson Tong at the Tencent Global Digital Ecosystem Conference on October 29. The release intends to integrate visual understanding into the company's core AI model ecosystem.
Building upon the Hy4 preview released in August, the upcoming full version is designed to improve performance in tasks involving sustained reasoning and practical productivity workflows. The company is also releasing Hy4 Flash, a version optimized for cost efficiency while retaining core features such as visual understanding and AI agent capabilities.
The Hy4 preview utilizes a architecture of 770 billion total parameters, with 49 billion activated during inference, supporting a context window of one million tokens. Tong stated that Tencent is adopting an integrated development strategy for its models and products to ensure performance improvements in practical, real-world applications.