Zhipu releases GLM-4.6 with a 200K context and domestic-chip support

Background

GLM-4.6 was released as an update to GLM-4.5. The release compared it with DeepSeek-V3.1-Terminus (深度求索 DeepSeek-V3.1-Terminus) and Claude Sonnet 4, two models of the time, to show where it stood.^[1]

What happened

Zhipu released GLM-4.6 in late September 2025. Its weights were posted to Hugging Face on 29 September 2025.^[2][3] The model is released under the MIT License, and its checkpoint has about 357 billion parameters (357B).^[1]

Key details

The context window grew from 128K to 200K tokens.^[1] Zhipu reported higher coding scores and better real-world coding, including front-end pages. The model also showed stronger reasoning with tool use during inference, better search agents, and more natural writing and role-play.^[1]

Zhipu evaluated the model on eight public benchmarks and said it was competitive with DeepSeek-V3.1-Terminus and Claude Sonnet 4.^[1] Hardware support covered FP8 and Int4 quantization on Cambricon (寒武纪) chips and native FP8 on Moore Threads (摩尔线程) GPUs.^[2]

Impact

Support for domestic chips means the model can be run in lower-precision formats on hardware made in China, which lowers the memory needed to serve it. The longer context window also lets the model handle longer inputs in one session.^[1][2]

Reading the benchmark figures

A benchmark score is only meaningful when two models are run on the same test under the same conditions. Vendors usually publish comparison tables that set their model against selected rivals on selected tests, so a number should be read together with the name of the test and the models it is compared with. A small gap between two scores does not by itself show a real difference in ability, and one test rarely captures everything a model can do.

Open weights and hosting

Open weights means the trained parameter files of a model are published so that others can download and run it. Open weights are not the same as open source, because whether a model counts as open source also depends on its licence, which decides what users may do with the weights. Hugging Face is a platform that hosts models and their weights for download, and several of the Zhipu models in this series were posted there.