June 2, 2026 — OpenAI has made its frontier AI models and Codex coding agent generally available on Amazon Web Services, removing what the company describes as one of the most persistent hurdles in enterprise AI deployment: getting cutting-edge models into production without overhauling an organization’s existing security and compliance infrastructure.
The move marks a significant deepening of the partnership between OpenAI and AWS, giving millions of businesses that already run their operations on Amazon’s cloud a direct path to building with OpenAI tools — through workflows, procurement systems, and governance frameworks they already have in place.
Specifically, two of OpenAI’s most capable models — GPT-5.5 and GPT-5.4 — are now accessible through Amazon Bedrock. According to OpenAI, GPT-5.5 is designed to handle the most demanding workloads, covering coding, advanced reasoning, agentic pipelines, and complex professional tasks. GPT-5.4 is positioned as the better value option, offering a strong price-to-performance ratio. Both models run on Bedrock’s next-generation inference engine and are accessible through the Responses API, with pricing that mirrors OpenAI’s own rates and no additional fees tacked on.
The integration means organizations can access these models the same way they would any other service on Bedrock — complete with IAM-based access management, AWS PrivateLink, encryption in transit and at rest, CloudTrail logging, and existing compliance controls. There is no new security model to learn and no separate infrastructure to set up. Usage also counts toward existing AWS cloud commitments, simplifying both procurement and financial oversight.
Codex, OpenAI’s AI-powered coding agent, is also now part of the Bedrock lineup. More than 4 million people use Codex weekly for tasks ranging from writing and refactoring code to debugging, explaining complex systems, and generating test coverage. Enterprise teams can now authenticate through their existing AWS credentials and run Codex inference through Bedrock’s infrastructure. It is available via the Bedrock API, the Codex CLI, the desktop app, and a Visual Studio Code extension.
A third offering — Amazon Bedrock Managed Agents powered by OpenAI — also launched alongside the two models. This service is aimed at teams looking to build production-grade AI agents without having to manually stitch together the underlying components. It combines OpenAI’s frontier reasoning capabilities with AWS infrastructure to handle persistent memory across sessions, skills that encode procedures, identity-based permissions enforcement, and flexible compute options. The goal is to let teams focus on defining what their agents should do rather than spending time engineering the stack underneath.
“By making OpenAI capabilities available within familiar AWS environments, organizations can spend less time navigating operational barriers and more time building,” OpenAI said in a statement accompanying the announcement.
The general availability follows a limited preview that launched in late April 2026. For enterprises that had been hesitant to adopt frontier AI due to concerns around compliance, vendor lock-in, or integration complexity, the AWS availability offers a route that works within cloud investments they have already made.
Leave a Reply
You must be logged in to post a comment.