Z.aiGLM Vision
GLM-4.6V
A vision-language release for documents, video and visual tools.
Practical use
Practical guidance. Preserve page images when layout carries meaning. Review charts, table headers and footnotes separately, and retain the image location behind every extracted value.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Downloadable |
| Model identifiers | glm-4.6vzai-org/GLM-4.6V |
| Context | 128,000 tokens during training.GLM-4.6V model card |
| Visual workflow | Can use multimodal tool inputs and interpret page images without first flattening them to text.GLM-4.6V model card |
| Licence | MIT.GLM-4.6V model card |
| Access | Current API catalogue and weights.Current Z.ai API catalogue |
Related models
A multimodal model built from GLM-4.5-Air.
Launch date not confirmedAPI · Downloadable- GLM-4.1V-9B-Thinking
Z.ai · GLM Vision
A visual reasoning model derived from GLM-4-9B-0414.
Launch date not confirmedDownloadable - GLM-4.6V-Flash
Z.ai · GLM Vision
The compact downloadable member of the GLM-4.6V line.
Launch date not confirmedAPI · Downloadable