GLM-4.7 is an open-source large language model from Z.ai designed to support real development workflows, emphasizing long-running task execution, stable tool calling, and multi-step reasoning.

The model is accessible through the BigModel.cn API and integrated into Z.ai's full-stack development platform, signaling broader adoption of open-source models in production environments.

Compared with GLM-4.6, GLM-4.7 delivers improved code generation, more capable complex reasoning, and more consistent, controllable performance on extended tasks with cleaner language output.

Benchmark results place GLM-4.7 among top open-source models, scoring highly on BrowseComp and τ2-Bench, and nearing Claude Sonnet 4.5 in coding-focused evaluations, while ranking first among open-source models in Code Arena’s large-scale blind evaluation.

In 100 production-like programming tasks spanning front-end, back-end, and command-execution scenarios, GLM-4.7 achieved higher task completion rates and greater stability than GLM-4.6 and is now the default model for Z.ai’s GLM Coding Plan.

Resources accompanying the release include Hugging Face model weights, a technical blog detailing GLM-4.7, and access options such as a default coding plan subscription and a try-it-now chat.