OpenAI Codex on AWS Bedrock offers enterprise controls, Digg reports

OpenAI and Amazon Web Services said on June 1 that Codex and OpenAI’s GPT-5.5 and GPT-5.4 models are now generally available on Amazon Bedrock, expanding an earlier preview into a production offering for AWS customers. AWS said the launch lets companies run OpenAI models and the Codex coding agent inside existing AWS security, billing and governance setups. OpenAI said enterprises can use the service through systems they already use for procurement, compliance and access control. Digg highlighted the move on June 2 in a post about enterprise controls around Codex workflows. ### What did AWS and OpenAI actually launch? Amazon Web Services said Codex, GPT-5.5 and GPT-5.4 are now available on Amazon Bedrock, its managed platform for foundation models. AWS said GPT-5.5 is aimed at harder coding and reasoning workloads, while GPT-5.4 is positioned on price-performance, and both can be called through the Bedrock Responses API. (aws.amazon.com) OpenAI said the AWS rollout gives customers “a new path” to use OpenAI through the cloud platform many enterprises already run. The company said the goal is to let businesses deploy frontier AI through existing security, compliance, procurement, billing and governance workflows rather than creating a separate operating model around OpenAI-hosted infrastructure. ### Where does Codex fit into this setup? (aws.amazon.com) AWS said Codex on Bedrock brings OpenAI’s coding agent into AWS environments where enterprise developers already work. According to AWS, Codex will be available through the Codex CLI, desktop app and VS Code extension, with inference routed through Bedrock. OpenAI’s developer documentation says Codex can be configured to use Amazon Bedrock as the model provider while the tool itself runs locally. (openai.com) In that setup, OpenAI said model requests go to Bedrock and “the OpenAI-hosted Responses API isn’t in the request path,” which shifts authentication and access control to AWS-managed layers. ### Which enterprise controls are AWS emphasizing? (aws.amazon.com) AWS said OpenAI models on Bedrock inherit controls customers already use, including AWS Identity and Access Management, AWS PrivateLink, encryption, guardrails and CloudTrail logging. AWS also said customers authenticate with AWS credentials and run inference through Bedrock, rather than through a separate OpenAI billing and identity stack. (developers.openai.com) Amazon said pricing for GPT-5.5 and GPT-5.4 matches OpenAI’s first-party rates and that usage counts toward AWS commitments. That matters for enterprise buyers because it ties model spending to existing AWS procurement and finance processes, according to Amazon’s June 1 announcement. ### Why did Digg frame this around CLI and IDE workflows? (aws.amazon.com) Digg’s June 2 item pointed to a practical enterprise use case: developers can keep using Codex in terminal and editor-based workflows while companies keep AWS-native identity, observability, billing and compliance controls around those requests. That framing matches AWS’s and OpenAI’s published documentation describing Codex access through the CLI and IDE integrations with Bedrock in the middle. (aboutamazon.com) AWS’s own description of the product says Codex is available through the CLI, desktop app and VS Code extension, while OpenAI’s documentation describes local execution with Bedrock handling model access. Taken together, the two companies’ materials support Digg’s description of Codex as fitting into enterprise developer workflows without requiring a separate control plane for identity and logging. That is an inference from the published product materials. (digg.com) ### What should companies watch next? April 28 was the date OpenAI and AWS first announced an expanded partnership bringing OpenAI models, Codex and managed agents to AWS in preview, and June 1 was the date both companies said general availability began. The next concrete reference points are the Bedrock product pages, AWS pricing details and OpenAI’s Codex-on-Bedrock documentation as enterprises test the CLI, desktop app and VS Code integrations in production. (aws.amazon.com) (openai.com)