Tencent’s Hy4 model gains in open-source AI rankings after ecosystem-driven training
Tencent Holdings' new Hy4 preview model has improved its standing in open-source AI rankings due to its training methodology. Analysts from Goldman Sachs noted that Tencent employed a "product-plus-model strategy," deploying preview models across its own product suite.

Briefing Summary
AI-generatedTencent Holdings' new Hy4 preview model has improved its standing in open-source AI rankings due to its training methodology. Analysts from Goldman Sachs noted that Tencent employed a "product-plus-model strategy," deploying preview models across its own product suite. This allowed Tencent to gather user data from real-world interactions, which was then incorporated into subsequent training rounds. This closed-loop approach is considered particularly effective for productivity and coding tasks, driving model differentiation in the development of AI agents. This strategy has helped bring Tencent's flagship model suite back into the top tier of open-source AI offerings.
Article analysis
Model · rule-basedKey claims
4 extractedPreview models were first deployed across Tencent’s suite of products, enabling user data collection for subsequent training.
Tencent’s Hy4 model gains in open-source AI rankings after ecosystem-driven training.
This closed-loop approach is particularly relevant for productivity and coding workloads where real-world task trajectories drive model differentiation.
Tencent’s strategy of using its vast product ecosystem to train its new Hy4 preview model gives it an edge in developing AI agents.