Introducing Opus 4.5 in Kiro
Adnan Ijaz
Product Lead
We’re excited to offer support for Claude Opus 4.5. Opus 4.5 is Anthropic's newest and most intelligent model, setting new standards across coding and agentic workflows.
Opus 4.5 achieves state-of-the-art performance on SWE-bench. It can better balance tradeoffs and ambiguity as it reasons across multiple systems to implement development tasks or fix complex bugs. Opus 4.5 is well-suited to power agents that can perform long duration tasks and reason across complex codebases, making it ideal for spec-driven development.
With this update, you can use Opus 4.5 to solve your most complex production software development tasks, generate detailed specs and architectural designs, while Auto can help you implement day-to-day features quickly and cost-effectively.
This model is available to users logging in with Google, GitHub, and AWS BuilderID in the Kiro IDE and the Kiro CLI. Opus 4.5 is not currently available for customers logging in with AWS IAM Identity Center. Opus 4.5 features a 2.2× credit multiplier, compared to 1.3× for Sonnet 4.5 and 0.4× for Haiku 4.5.
Opus 4.5 in Kiro is available with experimental support. We’re excited to see what you build and we’d love to hear your feedback on Discord or X.