Summary
On August 24, 2026, OpenAI made its GPT-5.6 family (Sol, Terra, and Luna) available in Kiro, AWS's spec-driven development environment. Joint testing on Terminal-Bench 2.1 showed the integration cut the cost of completed tasks by about 82%, across Kiro's plan-build-review-test workflow.
What changed
GPT-5.6 Sol, Terra, and Luna became selectable models in Kiro, AWS's spec-driven coding environment. The models cover Kiro's full workflow: turning requirements into structured implementation plans, executing multi-step coding tasks, reviewing output at checkpoints before changes land, and checking implementations with property-based testing. Joint Terminal-Bench 2.1 testing showed roughly 82% lower cost per completed task.
Why it matters
Putting OpenAI's frontier family inside AWS's spec-driven IDE widens where GPT-5.6 competes for coding workloads and gives Kiro users a cheaper path to completed tasks. It is another sign that model access is becoming multi-home across rival clouds and tools, with cost-per-completed-task emerging as the headline metric.
Evidence excerpt
OpenAI's GPT-5.6 family (Sol, Terra, Luna) is now available in Kiro, AWS's spec-driven development environment; joint testing on Terminal-Bench 2.1 showed the integration cut the cost of completed tasks by roughly 82%.