Models

Kiro gives you access to frontier and open weight AI models from OpenAI, Anthropic, and other providers. GPT-5.6 brings OpenAI models to Kiro for the first time, with three tiers that balance agentic performance and cost. Pick the right model for the job, or select Auto to let Kiro route each task to the optimal model automatically.

US and EU are geographies, each spanning several AWS Regions rather than a single Region. All authentication methods are supported for every model. For the Regions in each geography and the endpoint that serves your requests, see Inference endpoint regions.

Cost is relative to Auto (1.0x baseline). For example, a task that costs 10 credits on Auto would cost 22 credits on Opus, 4 credits on Haiku, or 0.5 credits on Qwen3 Coder Next.

The geography that serves your request depends on both the model you select and the region of your Kiro profile. GPT-5.6 models are served from the US regardless of your profile region. Every other model is served from the geography that matches your profile.

Profile region applies to enterprise users who sign in through IAM Identity Center or an external identity provider. Free Tier users and individual subscribers are always served from the US.

Kiro is powered by Amazon Bedrock, which uses cross-region inference to distribute requests across the Regions within a geography. Your request can be processed in any Region listed for your geography.

Cross-region inference does not change where your data is stored. Models marked experimental are an exception to the table above: they may be processed in commercial AWS Regions worldwide, including outside your profile's geography. See data protection for the full reference, and Amazon Bedrock cross-Region inference for how inference profiles route requests.

Model availability can vary by country or region. Kiro's model offerings align with each provider's usage and geographic requirements. For more information, see supported countries and regions: OpenAI, Anthropic, MiniMax, Zhipu AI (GLM), DeepSeek, and Qwen.

For models that support configurable reasoning effort, you can control how much reasoning the model applies to your prompts. Lower effort levels produce faster, shorter responses and use fewer credits. Higher levels spend more tokens on deeper analysis, multi-step reasoning, and thorough code generation.

Set the level from the model selector's Effort panel in the IDE, or with /effort (or the --effort launch flag) in the CLI. Your choice persists, and the picker only shows levels supported by your current model. See Reasoning effort for the full reference: per-surface mechanics, supported models, persistent per-model defaults, thinking behavior, and precedence.

• Start with Auto for most work. It optimizes both quality and cost automatically.
• Use GPT-5.6 Sol for the hardest multi-step work - long-horizon refactors and complex terminal tasks where you need state-of-the-art agentic performance.
• Use GPT-5.6 Terra or Luna when you want strong agentic capability with throughput or cost as the primary concern. Terra for balanced performance, Luna for maximum efficiency.
• Switch to Opus 5 when you hit a wall on a complex problem, need sustained multi-file work, or want the strongest multi-agent coordination and code review accuracy.
• Use Sonnet 5 when you want strong agentic behavior at a lower cost than Opus, especially for multi-step tasks that need to run to completion.
• Use Haiku for quick iterations, simple fixes, or when you want to conserve credits.
• Monitor your usage in your account settings to understand how model choice affects consumption.
• Factor model cost into your tier: If you primarily use Opus, consider Pro+, Pro Max, or Power for more credits. See plans and billing for details.

For detailed descriptions of each model's capabilities, strengths, and lifecycle status, see Available models.