AI
AWS adds Z.ai’s GLM-5.3 to Bedrock with a one-million-token context
AWS has made Z.ai’s GLM-5.3 available to eligible enterprise customers through Amazon Bedrock. The open-weight mixture-of-experts model offers a one-million-token context, up to 128,000 output tokens, prompt caching and integrations for agents, tools and structured outputs. Its improvements over GLM-5.2 are attributed to post-training rather than a new base model. The deployment is cross-Region only, which limits use where data must remain in one jurisdiction, and a licensing condition may require security review for very large companies reselling access.
Editorial summary · Publisher article
What matters
- GLM-5.3 has 744 billion total parameters and about 40 billion active per token.
- Bedrock supports prompt caching, Guardrails, Agents, Flows, function calling and structured JSON output.
- The model is available through cross-Region inference profiles rather than single-Region on-demand inference.
- Z.ai’s license requires companies with more than $10 billion in revenue to pass a security review before selling access.
Best-fit workloads
The deployment is aimed at coding agents, repository-scale refactoring, migration pipelines and long agent loops that can reuse large prompt prefixes. The source cautions that headline context limits do not replace task-level testing of tool success, latency, token use and cache behavior.
The main operational constraint
Applications requiring strict single-Region residency cannot use the current deployment model. Teams must also confirm account eligibility, permitted Regions and the licensing implications before adopting it in production.
Context & caveats
- The report says account eligibility must be confirmed and recommends production evaluations on complete repository tasks rather than relying on advertised context length.
Why this made the edition
Detailed account of a significant enterprise model deployment, including architecture, capabilities, licensing, routing limits and operational evaluation requirements.
Source material