GLM 5.3 hits Amazon Bedrock, targets coding and security

GLM 5 arrived on Amazon Bedrock earlier this year. The new GLM 5.3 builds on that lineage with stronger coding and emergent cyber security capabilities, and deeper Amazon Bedrock integration. Z.ai claims competitive performance on coding benchmarks including DeepSWE, Terminal Bench 3.0, and FrontierSWE, plus a 50% improvement over GLM 5.2 on its own internal coding benchmark. Direct comparisons to GLM 5 were not reported, because the magnitude of improvements led to updating the benchmark tests themselves since the GLM 5.1 announcement.

On Bedrock, GLM 5.3 supports implicit and explicit prompt caching, cross-Region inference profiles, and service tiers (Flex, Priority, Standard). It can be used with the open-source Strix penetration testing agent, and as of this writing the Strix documentation uses GLM 5.3 as its default model. The article frames this as complementary to AWS Continuum, a managed penetration testing service.

The outcome may hinge on whether the reported benchmark strengths translate into practical gains for enterprise coding and security teams, and on how quickly third-party tooling such as LiteLLM resolves the model alias so that configuration becomes simpler.