Zhipu releases GLM-5, a 744B open model

Background

GLM-5 scaled up from GLM-4.5, which has 355 billion total parameters with 32 billion active.^[1]

What happened

Zhipu released GLM-5 on 12 February 2026 under the MIT License.^[2][1] The release was described as an open model with 744 billion parameters, which is more than twice the size of GLM-4.5 in total parameters.^[1]

Key details

GLM-5 has 744 billion total parameters, with 40 billion active.^[1] Pre-training data grew from 23 trillion to 28.5 trillion tokens.^[1] The model integrates DeepSeek Sparse Attention (DSA) to cut deployment cost while keeping long-context capacity.^[1]

On SWE-bench Verified GLM-5 scored 77.8, and on SWE-bench Multilingual it scored 73.3. On Terminal-Bench 2.0 it scored 56.2 and 60.7. On HLE it scored 30.5, and 50.4 with tools. On τ²-Bench it scored 89.7, and on BrowseComp it scored 75.9 with context management.^[1]

Zhipu claimed best-in-class performance among open-source models on reasoning, coding and agentic tasks. Its technical report, "GLM-5: from Vibe Coding to Agentic Engineering", is arXiv 2602.15763.^[1]

Impact

The combination of a large parameter count and sparse attention is the main technical point of the release. DSA is meant to keep long-context work affordable to deploy, and the technical report documents the approach.^[1]

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.