Amazon Added Zhipu AI GLM-5.3 to Bedrock

The massive 744 billion parameter model is now accessible through AWS infrastructure for enterprise users.

Updated on Oct. 6, 2026 in Artificial Intelligence

Bold flat-color editorial illustration featuring a stylized geometric server module connected by an orange line, representing enterprise cloud infrastructure integration.
Amazon has integrated Zhipu AI's 744 billion parameter GLM-5.3 model into its Amazon Bedrock platform, providing enterprise clients immediate global access. AI Illustration. Upload story photo >

Live Poll

Do you trust US cloud platforms hosting AI models developed by foreign-based tech companies?

Amazon has integrated the Zhipu AI GLM-5.3 model into its Amazon Bedrock platform. This large language model release is available to enterprise customers immediately.

Why it matters

Enterprises can now utilize the model through Amazon Bedrock, allowing firms to bypass internal procurement hurdles linked to direct vendor contracts. The integration provides global access to high-parameter compute resources within established cloud workflows.

The GLM-5.3 model features 744 billion total parameters with 40 billion active parameters per query. It offers a 1 million token context window and a 128,000 token output limit, priced at $1.40 per million input tokens and $4.40 per million output tokens.

The players

Amazon

A global cloud computing provider operating the Bedrock platform for generative AI model deployment.

Zhipu AI

A deep learning research firm listed on the Hong Kong Stock Exchange that develops the GLM series of large language models.

The details

GLM-5.3 employs a Mixture-of-Experts architecture, a design that routes tasks through specific subsets of a neural network rather than activating the entire parameter set for every query. Users can access the model via US and global cross-Region inference profiles in Amazon Bedrock. The platform supports standard OpenAI-compatible Responses API as well as AWS native Invoke and Converse APIs.

Timeline

  1. January 2026: Zhipu AI made its Hong Kong stock debut.

  2. August 14, 2026: Zhipu AI released GLM-5.3 on its own API.

  3. August 28, 2026: GLM-5.3 model weights were released on Hugging Face.

  4. October 5, 2026: AWS added GLM-5.3 to Amazon Bedrock.

The Tech Race

This integration marks the transition of the GLM-5.3 model from research release to enterprise-grade availability on a major global cloud platform. It follows the model's initial debut on Zhipu AI's API and the public release of its weights on Hugging Face earlier this year.

Developers and enterprise users can now access GLM-5.3 directly through AWS Bedrock using existing Invoke or Converse APIs. Access is currently available via US and global inference profiles, allowing for immediate integration into workflows without individual vendor procurement.

The takeaway

The move underscores the growing trend of high-parameter models moving from open research weights to managed cloud availability. Watch for the forthcoming Shanghai STAR Market listing, which will provide further insight into the company's valuation after a 2,400% surge since January 2026.

What happens next

Zhipu AI has announced plans for a $2.2 billion follow-on listing on Shanghai's STAR Market, with a specific date yet to be disclosed.

Further reading

For more developments in large-scale model deployments, browse our Artificial Intelligence section.

Source note: This article includes information reported by Startup Fortune.

Live Poll

Do you trust US cloud platforms hosting AI models developed by foreign-based tech companies?