Tencent releases open-source Hy4 preview large language model

Tencent has released and open-sourced the Hy4 preview, a next-generation large language model featuring 770 billion total parameters and 49 billion active parameters. The model includes a context window exceeding 1 million tokens and is designed to enhance productivity in areas such as coding, office work, and scientific research.

Analysts from Goldman Sachs noted that Tencent’s strategy involves a “differentiated product-plus-model strategy,” utilizing its vast product ecosystem to collect user data and feed it back into training cycles. This closed-loop approach is seen as particularly effective for productivity and coding workloads.

In recent benchmarks, the Hy4 preview demonstrated significant improvements over its predecessor, Hy3. On the Code Arena WebDev leaderboard, the model ranked eighth globally, outperforming Alibaba Group Holding’s Qwen 3.8-Flash-Next and placing just behind Anthropic’s Claude Fable 5. The model is accessible through Tencent products like WorkBuddy and CodeBuddy, as well as via API through Tencent Cloud TokenHub and OpenRouter.