Claude Opus 5 is now available in Kiro

You just got a better Opus-class model for some of your hardest tasks in Kiro. Claude Opus 5 is available today across Kiro IDE, CLI, and Web — and it's a meaningful step up from Opus 4.8: stronger on the work that takes longest, better at coordinating parallel agents, and built with new controls that give you finer-grained handling of cost and latency in long sessions.

Opus 5 is a workhorse for the hardest agentic tasks. It excels at multi-file features, large refactors, and end-to-end feature work — completing full tasks rather than leaving stubs or placeholders.

In Kiro, that means spec-driven workflows in the IDE run to higher fidelity with less drift, and multi-step tasks finish more reliably with less supervision. For code review, Opus 5 finds a high rate of real bugs per pass with low false positives, and stays accurate at lower effort settings, useful for fast review passes at commit time and more thorough passes later. Opus 5 verifies its own work, iterates carefully until it succeeds, and completes full tasks rather than leaving stubs or placeholders.

We’ve made Opus 5 available in Kiro because Opus 5 is a strong agentic coding model built for long-running, multi-step work. It deeply understands your codebase, holds the thread across complex tasks, and pins down requirements for feature development and bug-fixing more effectively than Opus 4.8. We’re excited to see how developers will innovate with Opus 5.

—Deepak Singh, VP of Agentic AI at AWS

Opus 5 also introduces meaningful gains in multi-agent coordination: fewer cases of agents overwriting each other's work, and more effective use of writer-verifier patterns. If you run parallel agents in Kiro's CLI or IDE, Opus 5 is the model to test first.

Opus 5 ships with elevated cybersecurity safeguards. The classifiers allow Opus 5 to find vulnerabilities in source code, but block binary-based vulnerability scanning, penetration testing, and exploit generation. Anthropic's data shows the vast majority of sessions are unaffected. If you hit a false positive on a benign task, switch to another model in Kiro.

Claude Opus 5 is gradually rolling out with experimental support to Kiro Pro, Pro+, Pro Max, and Power customers in us-east-1 (IAD) and eu-central-1 (FRA), with cross-region inference support. It ships with a full 1M context window and a 2.2x credit multiplier.

Download Kiro or restart the IDE or CLI to check for the latest available models. Web users can refresh their browser to see Claude Opus 5 in the model selector.