Is this page helpful?

Release Notes#

Release 2.1.2-variant#

This release contains the following model update.

GLM-5.3-Flash#

This is the initial release of GLM-5.3-Flash. For more information about this model, refer to the model card.

Follow the instructions in the Get Started with GLM-5.3-Flash guide to deploy the model. For two-node NVIDIA DGX Spark deployment, refer to Deploy on DGX Spark.

For GPU support, refer to the support matrix for GLM-5.3-Flash.

Note the following limitation:

  • Video inference requires an external FFmpeg 8 runtime with libavcodec.so.62 mounted at /opt/ffmpeg8.

For information about past updates and older versions, refer to the previous release notes.