GPT-5.5, GPT-5.4, and Codex are now generally available on Amazon Bedrock. Deploy them to your production applications and agents today using Bedrock’s high-performance inference engine.
Important points
- OpenAI’s most advanced frontier model, GPT-5.5, is generally available on Amazon Bedrock. Pricing matches OpenAI’s first-party pricing.
- Amazon Bedrock’s Codex is generally available with a pay-per-token fee. Inference is performed through Bedrock and usage counts against your existing AWS commitments.
every monthafter us Announcement of expanded partnershipGPT-5.5, GPT-5.4, and Codex are now generally available. Amazon bedrock, This gives you access to frontier models and OpenAI coding agents for software development.
Amazon Bedrock is A platform for building and running AI applications and agents at production scale. OpenAI models on Bedrock Runs on Amazon Bedrock’s next-generation inference engine, built for high performance, reliability, and security.
The most capable OpenAI models on Amazon Bedrock
GPT-5.5 understands user intent faster, handles multi-step tasks autonomously, and excels at writing and debugging code across large code bases, analyzing data, generating documents and spreadsheets, and interacting with software across multiple tools until tasks are completed. This improvement is most noticeable in agent coding and knowledge work, where real progress depends on maintaining context and taking action over time.
Both GPT-5.5 and GP-5.4 are built for complex, multi-step tasks and are currently available in the Amazon Bedrock model catalog. You can call them through Amazon Bedrock’s Responses API and pay the same per-token fee as directly from OpenAI at no additional charge.
Bedrock’s inference engine It provides unique, isolated queues with automated capacity management, resulting in predictable performance even under high loads. As each request executes, its complete state is persistently and continuously captured, so if hardware fails or a node reboots mid-call, the request picks up where it left off, rather than starting over. All calls are made using the governance controls you already use across AWS (IAM permissions, VPC, and private link Isolation, KMS encryption, AWS CloudTrail audit logs. Prompts and responses are not used to train the model or shared with model providers. These protections extend to GPT-5.5 and GPT-5.4 on Amazon Bedrock.
“At Amgen, we are focused on applying advanced AI in ways that help accelerate the delivery of potential new treatments while equipping teams with advanced tools. OpenAI’s GPT-5.5 and Frontier models deliver compelling advances in functionality, quality, and consistency that are important in fields where problems are complex and standards for scientific accuracy and decision quality are very high. By making these models available on AWS, we are committed to responsible AI. It gives you significant new avenues to explore and extend these capabilities within your enterprise-wide security, governance, and operational framework.”
– Sean Bruich, Senior Vice President and Chief Technology Officer, Amgen
Accelerate software development with Codex on Amazon Bedrock
Codex is an OpenAI coding agent for AI-powered software development. more 5 million people Use Codex every week to write, refactor, debug, test, and validate code across large codebases. Codex maintains the context of the entire repository, reasoning through vague failures, using tools to check assumptions, and make changes to the surrounding code with an awareness of how the system is connected. With GPT-5.5 inference, Codex completes the same work more efficiently and with higher quality compared to previous model versions.
Amazon Bedrock Codex It is available through IDE integration with Codex App, Codex CLI, Visual Studio Code, JetBrains, and Xcode, and all model inference is routed through Amazon Bedrock. Inference remains within the selected region to meet data residency requirements. You pay per token, so there are no seat licenses or per-developer contracts. So you can get started right away and expand your access as you continue using it.
“Autodesk is a technology platform for the people who design and build the world around us. Workflows like architectural design are highly iterative and require precision, coordination, and continuous improvement across teams. With the general availability of OpenAI models and Codex on Amazon Bedrock, our team is able to leverage frontier AI capabilities and AI on a scalable and secure AWS infrastructure. We are evaluating how our development tools can accelerate development workflows and help our customers make more informed decisions.”
– Ritesh Bansal, Vice President of Analytic Data, Agentic AI, and AI/ML Platforms at Autodesk.
what’s next
In our expanded partnership announcement, we introduced Amazon Bedrock managed agent powered by OpenAI. You will soon be able to deploy production-ready agents built on the OpenAI Agent Harness, delivering faster execution times, sharper inference, and reliable steering of long-running tasks. Every agent runs under a unique identity and logs all actions to make them auditable and run on every model inference on Amazon Bedrock. To stay up to date, sign up at interest form.
We continue to expand the OpenAI capabilities available on Amazon Bedrock as new advances arrive. This includes Daybreak, an OpenAI vision to change the way software is built and defended. Daybreak includes Cyber Models and Codex Security, designed to help cyber defenders identify vulnerabilities, see code risks, and guide remediation throughout the development lifecycle. Once Daybreak is available on Bedrock, security teams will be able to deploy it through the governance and operational frameworks they already use on AWS.
Let’s get started
GPT-5.5 and GPT-5.4 are currently available in the Amazon Bedrock model catalog. CTogether AWS Region Page for availability. For documentation and step-by-step tutorials, please visit Amazon Bedrock Documentation and start a blog.
About the author
