Open models
Tencent posts Hy4’s open weights, and they need about 944 GB of GPU memory
Tencent says its 770B Hy4 reads images, drives a computer and writes game code. The weights are free to download, but running them takes an 8-GPU server.

Open models
Tencent says its 770B Hy4 reads images, drives a computer and writes game code. The weights are free to download, but running them takes an 8-GPU server.

Language models
Alibaba's Qwen4-Max is now on pay-as-you-go API at $1.50 in and $4.50 out. The efficiency claims are Alibaba's own; the first outside scores put it mid-pack.

Language models
Meituan's 1.6 trillion parameter model posts coding and agent scores near Claude Opus 5. Every score so far is Meituan's own, and the weights are not out.

Open models
Mistral says its 1-trillion-parameter model leads open-weight coding. Artificial Analysis ranks it second and measured lower scores on two of Mistral's headline tests.
