Haiku 5.5 now available on Kiro in AWS GovCloud (US)
AWS GovCloud (US) now offers Anthropic’s Haiku 5.5 model in the Kiro IDE and CLI, expanding the Claude 5.5 family with faster, volume-optimized performance and parallel review capabilities.
Amazon Web Services has introduced Haiku 5.5, the latest model in the Claude 5.5 family, now available in the Kiro integrated development environment and command-line interface within AWS GovCloud (US) Regions. This release completes the availability of all three Claude 5.5 models—Haiku, Sonnet, and Opus—within Kiro for government workloads. Haiku 5.5 is positioned as the fastest model in the family, designed for high-volume tasks requiring rapid responses such as summarization, classification, and bulk text processing.
The new model supports adaptive thinking across five effort levels, with medium set as the default, and includes a 1 million token context window. Haiku 5.5 operates with a 0.06x credit multiplier, while requests exceeding 100,000 tokens are billed at a 0.3x multiplier, reflecting its optimized cost structure for large-scale operations. Developers can deploy it as a parallel review subagent to assist with smaller edits or routing tasks while heavier planning and coding workloads are handled by Opus 5.5 or Sonnet 5.5.
To access Haiku 5.5, users must update their Kiro IDE or CLI to the latest version and restart the application before selecting the model from the interface. AWS provides documentation specific to GovCloud environments, monitoring guides, and direct support through account teams to assist with deployment and usage. The company also directs users to the Kiro product page for additional information about the environment and its capabilities.
The announcement follows AWS’s expansion of the Claude 5.5 family within its GovCloud infrastructure, reinforcing its commitment to supporting government and regulated workloads with specialized tooling. Haiku 5.5 joins Sonnet 5.5 and Opus 5.5 in Kiro, offering agencies and developers a tiered model suite for balancing speed, cost, and complexity in AI-assisted development tasks.