OpenAI GPT-5.5 and Codex Are Now on Amazon Bedrock — With AWS Security and Pay-Per-Token Pricing
Cloud Platform EngineeringOpenAI’s two most capable current models — GPT-5.5 and GPT-5.4 — are now generally available on Amazon Bedrock, alongside Codex, the AI coding agent serving more than four million developers weekly. The access model removes the friction that previously forced AWS-first organizations to maintain a separate OpenAI API relationship.
What the Bedrock Integration Actually Provides
The models arrive with Bedrock’s full security stack: IAM-based access control, VPC isolation, and encryption in transit and at rest. All inference runs inside the selected AWS Region — data doesn’t leave the region boundary, which closes the data residency objection that has kept regulated industries from deploying frontier OpenAI models.
Pricing matches OpenAI’s first-party rates with no added Bedrock fees, and usage counts toward existing AWS enterprise contracts — eliminating the separate OpenAI spend line from the vendor management picture. The model is pay-per-token with no seat licenses or per-developer commitments, which fits the variable inference patterns of agentic workloads better than flat-rate enterprise licensing.
The Model Capabilities and Access Path
GPT-5.5 is positioned for demanding workloads: complex reasoning, long-horizon agentic tasks, and multi-step coding workflows. GPT-5.4 offers the best price-performance balance — effectively the production default for workloads that don’t require GPT-5.5’s peak capability. Both models support tool orchestration, function calling, and multi-turn conversational state.
Codex integrates with VS Code, JetBrains, Xcode, and a standalone CLI, routing all inference through Bedrock’s infrastructure. For organizations that have already standardized on AWS-native AI infrastructure, this means Codex access without a second vendor relationship or compliance exception.
GPT-5.5 launched in one US region; GPT-5.4 is available in two US regions including GovCloud — the government-adjacent deployment path that has been a blocking issue for federal and defense customers evaluating OpenAI models.
The So What
The practical unlock for AWS platform teams is consolidation: frontier OpenAI models can now be provisioned, billed, and secured through existing Bedrock infrastructure without separate vendor agreements or data governance exceptions. The regional limitation (US-only at launch) will expand, but for organizations headquartered in the US with data residency requirements, it’s sufficient today. The more significant shift is competitive: Bedrock now carries both Anthropic Claude and OpenAI GPT-5.x, letting enterprise buyers compare models on identical infrastructure rather than inferring performance differences from separate API relationships with different security postures.
Source: AWS Blog
Content created with AI assistance and reviewed for accuracy.
Join the conversation
Stack Insiders is our free community for readers who want to go deeper — share resources, ask questions, and connect with others across every vertical we cover.
Join Stack Insiders →