GLM-5.3 is a public endpoint. New PyTorch templates go up to 2.13.͏‌ ͏‌ ͏‌ ͏‌   ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­  
Runpod Log in
GLM-5.3, now a public endpoint
GLM-5.3, now a public endpoint

GLM-5.3, Z.ai's model for coding and multi-step tasks, is now available through Runpod's API. Runpod hosts the model, so you can send requests with your Runpod API key without deploying it yourself. Pricing is $0.77 per million input tokens, $0.143 per million cached input tokens, and $2.42 per million output tokens. The getting-started guide covers trying it in the Playground and making your first API request.

Try GLM-5.3
What we published
Rebuilding our PyTorch templates
Rebuilding our PyTorch templates

PyTorch 2.6 through 2.13 across CUDA and Ubuntu versions, with a new CUDA dropdown when you deploy a Pod.

Explore templates
 
Terraform provider 2.0 is GA
Terraform provider 2.0 is GA

Manage Pods, Serverless endpoints, Instant Clusters, network volumes, and templates from config files.

Read the docs
Guide: Wan 2.2 in ComfyUI

Generate video from text or images, then turn your workflow into a Serverless API that scales to zero.

Read guide
 
Build Runpod templates with your coding agent

A new skill turns an app or GitHub repo into a template project for Pods or Serverless.

View the skill
Webinar: State of AI Compute, October 29
Photo of Sarah Melancon and Brooke Spath

Which models are teams actually running, and how much GPU memory do their workloads need? Sarah and Brooke walk through six trends from what 1.2 million developers deployed, and what each one means for how you size and spend. Join live on October 29 at 6 AM or 11 AM PT.

Save your seat
Meet us this week
ODSC West: Burlingame, Oct. 27–29

Learn more →

ComfyUI hackathon: San Francisco, Oct. 28

Learn more →

GitHub Hugging Face Discord X LinkedIn
Runpod Inc., 329 Bryant St #4D, San Francisco, CA 94107, United States
 
Terms and Conditions Preference Center Unsubscribe Contact Us
© 2026 Runpod, Inc. All rights Reserved.