OpenAI's frontier models and Codex coding agent are now generally available on Amazon Bedrock, one month after OpenAI revised its exclusive cloud arrangement with Microsoft to allow distribution through other providers. GPT-5.5 and GPT-5.4 are accessible through the Responses API on Bedrock's next-generation inference engine, with pricing matching OpenAI's direct rates and usage counting toward existing AWS commitments. More than 100,000 organizations already on Bedrock can now access OpenAI models without introducing a new vendor relationship, billing path, or governance process.
AWS VP Swami Sivasubramanian confirmed that GPT-5.5 is "available natively on Amazon Bedrock," alongside the full lineup of OpenAI models. OpenAI described the launch as the beginning of a broader expansion:
This is the start of a broader path for customers to bring frontier AI into the environments where they already build, govern, and ship.
The enterprise governance integration is the most significant aspect for teams evaluating this. Every OpenAI API call routed through Bedrock inherits AWS-native controls: IAM for identity and access, VPC and PrivateLink for network isolation, KMS encryption for data at rest, and CloudTrail for audit logging. OpenAI states that customer data is not used for model training and is not shared with model providers. Bedrock offers three inference routing options: In-Region for strict compliance, Geo Cross-Region for higher throughput within a geography (US or EU), and Global Cross-Region for maximum throughput without residency constraints.
On Hacker News, practitioners made clear why the Bedrock path matters more than the models themselves. One engineering manager at a large enterprise explained:
I work at a large enterprise with strict data governance built into customer contracts. Using vendors where you not only have infosec permission, but they are also listed as data processors in our contracts with our customers is the way not to get fired and sued.
Another commenter clarified the isolation model that makes this possible:
AWS Bedrock is other companies' models running on separate dedicated AWS hardware. If you use Codex via Bedrock, OpenAI never sees your data or prompts because they stay sandboxed in an ephemeral Bedrock instance. Over the past year, Claude being available via Bedrock and ChatGPT/Codex not being available via Bedrock has been a huge competitive advantage for Anthropic in the enterprise space.
Not everyone is convinced the governance story is complete. Marek Porycki, a strategic advisor, noted on LinkedIn that infrastructure controls and decision-level governance are not the same thing:
CloudTrail logs the API call. It does not log the decision rights. That gap is exactly where most pilots stall before production: technically green, accountability undefined.
For teams deploying agentic workflows where models act autonomously, knowing who called the model is different from knowing whether the action was authorized.
Codex, OpenAI's coding agent used by more than 5 million developers weekly, is available through the Codex App, CLI, and IDE integrations for Visual Studio Code, JetBrains, and Xcode, with all inference routed through Bedrock. The pricing model shifts from per-seat licensing to pay-per-token billing with no seat fees, a significant cost structure change for large developer teams. Codex on Bedrock uses GPT-5.5 for inference and inherits the same security controls as direct model calls.
GPT-5.5 is positioned as OpenAI's most capable model for agentic coding, data analysis, and autonomous multi-step tasks. GPT-5.4 targets the best price-performance ratio for production workloads at scale. GPT-5.5 is available in US East (Ohio). GPT-5.4 is available in US East (Ohio), US West (Oregon), and as of June 3, AWS GovCloud (US-West), making it the first OpenAI model available in a government cloud environment.
The competitive landscape on Bedrock has shifted considerably. AWS now hosts both Anthropic's Claude (backed by a multi-billion dollar AWS investment) and OpenAI's GPT models on the same managed infrastructure. Microsoft, which retains its investment in OpenAI, has simultaneously deepened its collaboration with Anthropic on Azure. The era of exclusive cloud partnerships for frontier AI models appears to be ending, replaced by a model where enterprises choose providers based on workload fit rather than cloud allegiance. One Hacker News commenter put the Anthropic implication bluntly:
Anthropic's incredible revenue run-up was basically a result of botched Gemini releases and OpenAI having their hands tied behind their Azure backs. Anthropic models were quite literally the only viable serverless API models on AWS.
Amgen CTO Sean Bruich and Autodesk VP Ritesh Bansal are among the cited early adopters. OpenAI also previewed Daybreak, a forthcoming capability bringing cyber models and Codex Security to AWS for secure code review, threat modeling, and dependency risk analysis. Teams should note that GPT models on Bedrock currently support only the Responses API, with console support coming later.
OpenAI models and Codex on Bedrock are available now. The OpenAI on Amazon Bedrock documentation provides setup guides and code examples.