「Introducing GLM 5」 — GLM 5.3 arrives on Amazon Bedrock with coding-agent and security-testing workflows

AWS says GLM 5.3 from Z.ai (Zhipu AI) is now available to eligible enterprise customers on Amazon Bedrock. AWS describes it as a 753-billion-parameter mixture-of-experts model optimized for coding and long-horizon agentic tasks. Customers access it through fully managed APIs rather than managing inference infrastructure themselves.

The model can be invoked through the OpenAI-compatible Responses and Chat Completions APIs, or through Bedrock’s Invoke and Converse APIs. AWS says implicit prompt caching is enabled by default, with explicit cache controls available on the first two APIs; each explicit cache breakpoint must contain at least 1,024 tokens to qualify. AWS also lists Flex, Priority, and Standard service tiers, as well as US and Global cross-Region inference profiles.

The AWS walkthrough also demonstrates an authorized security test of a customer’s own application using Strix, an open-source AI penetration-testing agent. For benchmark context, Z.ai’s own GLM-5.3 overview reports Terminal-Bench 3.0 results increasing from 4.6 to 28.3 and DeepSWE v1.1 results from 46.2 to 66.9; those figures are results reported by the model developer.