Tencent open-sources its Hy4 preview model

Tencent just put a 770-billion-parameter model on the internet for anyone to download. The catch is that Hy4 preview, the latest release from the Hunyuan team, is a Mixture-of-Experts system, so it only wakes up about 49 billion of those parameters for each token. The company said as much in an August 28 post, and it put the context window at more than 1 million tokens.

The weights are out under Apache License 2.0 on Hugging Face, ModelScope, GitCode, and CNB. The GitHub model card describes 78 layers, 256 routed experts plus one shared expert, and the top-8 routed experts activated per token. That sparse routing is why a 770B-total model can run while touching only 49B parameters on any given pass.

Hy4 preview is also wired into Tencent's own WorkBuddy, CodeBuddy, Yuanbao, and ima apps, with API access through Tencent Cloud TokenHub and OpenRouter. Tencent lists TokenHub prices at $0.834 per million input tokens, $2.501 per million output, and $0.042 per million cache-hit tokens. The model is free on WorkBuddy and CodeBuddy for two weeks from launch, and free access to the older Hy3 on both apps runs until September 30.

Tencent quotes an internal blind evaluation, 163 experts scoring 203 engineering tasks, in which Hy4 preview scored 2.99 out of 4.00 against GLM-5.3 at 2.92 and Kimi K3 at 2.94. Those are vendor numbers with narrow margins, not an independent bench and not a claim that the model beats GPT-5. The company says Hy4 preview "was expanded significantly in model size, context length, and data volume," and it reports a 31.8% gain in end-to-end inference throughput over its baseline. A further batch in the Hy4 series is described as expected soon. TechNode carried the same figures.

The open weights sit in the same crowded model market as OpenAI winding down its Cursor model contract.

Some offers on this page may be paid placements or contain affiliate links.