--- title: Amazon Bedrock Now Offers OpenAI Models, Codex, and Managed Agents: What It Means for Enterprise AI description: AWS just made OpenAI's frontier models, Codex, and production-ready Managed Agents available inside Amazon Bedrock, wrapped in IAM, PrivateLink, Guardrails, and CloudTrail. This is what changes for CTOs evaluating OpenAI direct vs. AWS. url: https://www.joinhorizon.co/blog/amazon-bedrock-openai-models-codex-managed-agents/ datePublished: 2026-04-30T00:00:00.000Z dateModified: 2026-10-28T00:00:00.000Z author: palaniappan-p category: AI Agents tags: amazon-bedrock, openai, codex, managed-agents, bedrock-agentcore, generative-ai, aws, enterprise-ai --- # Amazon Bedrock Now Offers OpenAI Models, Codex, and Managed Agents: What It Means for Enterprise AI > AWS just made OpenAI's frontier models, Codex, and production-ready Managed Agents available inside Amazon Bedrock, wrapped in IAM, PrivateLink, Guardrails, and CloudTrail. This is what changes for CTOs evaluating OpenAI direct vs. AWS. > **AWS lifecycle notice (June 30, 2026):** Amazon Bedrock Agents is now **Bedrock Agents Classic**, in maintenance for new customers after **July 30, 2026**. Net-new agent builds should use [Bedrock AgentCore](https://www.joinhorizon.co/blog/amazon-bedrock-agentcore-production/). Full matrix: [lifecycle roundup](https://www.joinhorizon.co/blog/aws-service-lifecycle-updates-june-2026/). On **April 28, 2026** AWS [announced](https://aws.amazon.com/about-aws/whats-new/2026/04/bedrock-openai-models-codex-managed-agents/) that Amazon Bedrock now offers OpenAI's latest models, the Codex coding agent, and a new class of Bedrock **Managed Agents** powered by OpenAI, all in Limited Preview. The [accompanying partnership announcement](https://www.aboutamazon.com/news/aws/bedrock-openai-models) frames it as bringing "frontier intelligence to the infrastructure you already trust." For CTOs and platform teams that have been running parallel workloads on OpenAI's API and Bedrock, the procurement story, the security story, and the agent-runtime story all collapse into a single AWS-governed surface. This post covers what was announced, what changes, and what to do this quarter, written from the perspective of an AWS Select Tier Services Partner deploying production GenAI for healthcare, SaaS, and ecommerce clients. > **Update (July 30, 2026):** GPT-5.6 Sol, Terra, and Luna are GA on Bedrock Mantle. AWS cut on-demand prices **80% for Luna** and **20% for Terra** (Sol unchanged). Routing and rate-card math: [GPT-5.6 Luna & Terra Bedrock price cuts](https://www.joinhorizon.co/blog/openai-gpt-5-6-luna-terra-bedrock-pricing-2026/). > **Update (September 22, 2026):** GPT-6 Sol and GPT-6 Luna are GA on Bedrock, alongside GPT-6 Astra (September 8, 2026). Routing and model IDs: [GPT-6 Sol and Luna on Bedrock](https://www.joinhorizon.co/blog/openai-gpt-6-sol-luna-amazon-bedrock-2026/). GPT-6.1 Sol followed on September 29, 2026 (Standard US Geo $2.20 / $11.00 per 1M tokens; the model card also lists Global CRIS global.openai.gpt-6.1-sol). Ultrafast followed on October 8, 2026 at six times Standard ($13.20 / $66.00 on that same short-context US Geo tier). > **Update (September 29, 2026):** Bedrock Managed Agents, powered by OpenAI, is now in public preview in us-east-1, us-east-2, and us-west-2. Tools run on **your** compute, either self-hosted or on AgentCore Runtime, not on AWS-owned compute. What AWS runs vs what you own: [Bedrock Managed Agents preview guide](https://www.joinhorizon.co/blog/amazon-bedrock-managed-agents-openai-preview-2026/). > **Note on Limited Preview (historical):** When this post first shipped (April 2026), specific OpenAI SKUs and on-demand pricing were still preview-gated. That catalog has since firmed up. Use the July 2026 pricing post and the live [Bedrock pricing](https://aws.amazon.com/bedrock/pricing/) page for current rates. ## TL;DR for CTOs - OpenAI's frontier models are now first-class citizens inside Amazon Bedrock, alongside Claude (including the June 2026 [Claude Fable 5 Mythos-class launch](https://www.joinhorizon.co/blog/claude-fable-5-aws-bedrock-mythos-class-guide-2026/)), Llama, Mistral, and Nova. - **Codex**, the coding agent used weekly by 4M+ developers, runs on Bedrock through the same CLI, desktop app, and VS Code extension developers already know, but authenticates with **AWS credentials** instead of personal OpenAI keys. - **Bedrock Managed Agents, powered by OpenAI** (public preview since September 29, 2026): AWS runs the conversation and model calls, while tools and commands run on compute you provide, either your own host or AgentCore Runtime. The preview has no built-in long-term memory. - Enterprise controls apply: **IAM, PrivateLink, Guardrails, encryption, CloudTrail**, and each agent runs with its own identity and audit log. - OpenAI usage **counts toward existing AWS cloud commitments**, so OpenAI spend rolls into the AWS bill rather than a parallel procurement. - The "Bedrock vs. OpenAI API" decision matrix, including [our existing comparison](https://www.joinhorizon.co/blog/aws-bedrock-vs-openai-api-enterprise/), has shifted. For most regulated enterprises, Bedrock is now the default path _to_ OpenAI, not an alternative _to_ OpenAI. ## What AWS actually announced (April 28, 2026) The announcement bundles three offerings that arrive together but solve different problems: ### 1. OpenAI frontier models on the Bedrock API OpenAI's latest models are accessible through the Bedrock `InvokeModel` and Converse APIs alongside Anthropic Claude 4.6, Meta Llama 4, Amazon Nova, Mistral, DeepSeek, Qwen, and the rest of the [~100-model Bedrock catalog](https://www.joinhorizon.co/blog/why-aws-bedrock-is-the-fastest-path-to-enterprise-genai/). For application teams, this means you can swap providers behind the same SDK call, and Bedrock-native features like [Knowledge Bases for RAG](https://www.joinhorizon.co/blog/how-to-build-rag-pipeline-amazon-bedrock-knowledge-bases/), [Guardrails](https://www.joinhorizon.co/blog/how-to-set-up-amazon-bedrock-guardrails-production/), and [Flows for orchestration](https://www.joinhorizon.co/blog/amazon-bedrock-flows-workflow-orchestration/) compose with OpenAI models the same way they compose with Claude. ### 2. Codex on AWS credentials Codex, OpenAI's coding agent, is available through the Bedrock API via three surfaces: - **Codex CLI** for terminal-based agentic coding sessions - **Codex Desktop App** for a longer-running workspace - **Codex VS Code extension** for inline integration The headline change is authentication. Developers sign in with **AWS credentials** (typically through IAM Identity Center / SSO), so each Codex session is attributable to a named human, scoped by IAM permissions, networked through your VPC, and logged to CloudTrail. The shadow-IT problem of personal `OPENAI_API_KEY` values in dotfiles goes away. ### 3. Bedrock Managed Agents, powered by OpenAI Of the three, this is the most significant. AWS's April announcement described a fully managed agent runtime where the agent itself (model + framework + memory + tool routing) is operated by AWS on top of Amazon Bedrock AgentCore. According to that announcement, Managed Agents include: - **Persistent memory** across sessions - **Skills** that encode procedures - **An identity** that enforces permissions (so the agent inherits and respects IAM) - **Compute options** suited to the task (default: AgentCore) The September 29, 2026 public preview splits the work differently: tools run on your compute, and there is no built-in long-term memory. See the update at the top of this post. Box CTO Ben Kus, the only named partner in the announcement, describes the value as: > "With Amazon Bedrock Managed Agents, powered by OpenAI, developers can build optimized, production-scale AI applications that bring together the strengths and capabilities of OpenAI's latest models with the scale, security, and infrastructure of AWS." For teams that have already shipped a Bedrock Agent using Claude, the comparison is direct: Managed Agents take you one rung higher in abstraction. You give up some control over the agent loop in exchange for AWS owning the runtime. ## OpenAI frontier models on Bedrock: what's different from the direct API If your team already calls OpenAI's API in production, here is what changes when the same call goes through Bedrock instead: | Concern | OpenAI direct | OpenAI on Bedrock | | --------------------- | ----------------------------------- | -------------------------------------------------------------------- | | Authentication | Per-developer or per-app API key | AWS IAM, IAM Identity Center, IAM Roles | | Network path | Public internet to `api.openai.com` | VPC + AWS PrivateLink option | | Compliance posture | OpenAI's own attestations | Inherits Bedrock's HIPAA-eligible, SOC 2, FedRAMP-in-process posture | | Audit logging | OpenAI dashboard / your own logging | CloudTrail, native | | Content safety | OpenAI moderation API (separate) | Bedrock Guardrails (composable) | | Billing | Separate OpenAI invoice | AWS bill, applies to commitments | | Model lifecycle | OpenAI deprecation calendar | OpenAI deprecation surfaced through AWS region/SKU lifecycle | | Region/data residency | OpenAI's regions | AWS regions (subject to preview availability) | The compliance and billing rows typically tip the decision. A healthcare customer running a HIPAA-eligible workload could not put OpenAI's direct API in the architecture without a separate BAA conversation. The same model, accessed through Bedrock, sits inside the AWS BAA they already signed. ## Codex on AWS credentials: what dev teams gain ![Architecture diagram of a developer signing in through IAM Identity Center to Codex, with the session logged in CloudTrail](../../assets/images/blog/amazon-bedrock-openai-models-codex-managed-agents-identity.webp) _Figure: Codex authentication in the article. The developer signs in through IAM Identity Center, and CloudTrail attributes the session. A personal API key is the path this replaces. The managed-agent turn is a different article. Model ids are not shown. [Open the draw.io source](https://www.joinhorizon.co/examples/architecture-blog-2026/amazon-bedrock-openai-models-codex-managed-agents/identity.drawio)._ Codex through Bedrock is the same Codex developers have been using (same CLI, same VS Code extension, same agentic coding behavior) wired to a different identity provider. Three improvements show up immediately: **1. Identity and offboarding actually work.** When a developer leaves, you remove their IAM Identity Center user and their Codex access goes with it, with no hunting for stranded API keys in `~/.config`. New hires get Codex through the same SSO group membership that already grants them access to AWS Console, S3, and the rest of the toolchain. **2. Per-developer audit trail.** Every Codex invocation can be attributed to a human via CloudTrail. For regulated environments, especially in financial services, this is the difference between "we use AI coding assistants" and "we use AI coding assistants and can prove who did what." **3. Network and data path control.** Codex traffic can flow over AWS PrivateLink. For customers with policies against sending source code over the public internet to model providers, this changes what is allowed. The trade-off is that you lose the personal-key escape hatch developers have been using to ignore corporate restrictions. ## Bedrock Managed Agents, powered by OpenAI Managed Agents sit one level above the [DIY Bedrock Agent pattern we documented earlier this month](https://www.joinhorizon.co/blog/how-to-build-amazon-bedrock-agent-tool-use-2026/). The DIY pattern gives you control: you define action groups, write Lambda tool implementations, set instructions, and own the agent loop. Managed Agents, by contrast, deliver the agent runtime as a service: AWS hosts and scales the loop on AgentCore, and you focus on tools, skills, and memory configuration. In practice: - **AgentCore** provides the default compute environment and is positioned as the open platform for "constructing, linking, and scaling agents across any model and framework." - **Persistent memory** survives session boundaries, so the agent remembers context across days and not just inside a single chat. - **Skills** encode procedures the agent can invoke, similar to action groups but framework-managed. - **Identity** is enforced through IAM, so an agent's blast radius is bounded by the IAM role it runs under, not by trust in the model. When does this beat building it yourself with Bedrock Agents Classic + Claude? We expect three patterns to dominate: 1. **Coding assistants beyond Codex.** Internal copilots that go past inline completion: agents that take a Jira ticket, branch the repo, draft a PR, and request review. 2. **Long-running enterprise workflows.** Multi-day processes (procurement, onboarding, claims triage) where session-spanning memory matters more than the per-call latency. 3. **Cross-system orchestration.** Agents that coordinate across Salesforce, ServiceNow, and S3 where AWS owning the runtime simplifies the operational story. For agents that need tight control over the loop (deterministic guardrails, custom routing logic, very specific tool sequencing), the [DIY Bedrock Agents Classic path](https://www.joinhorizon.co/blog/how-to-build-amazon-bedrock-agent-tool-use-2026/) and [Bedrock Flows](https://www.joinhorizon.co/blog/amazon-bedrock-flows-workflow-orchestration/) remain the right tools. ## Enterprise controls ![Architecture diagram of Codex reaching Bedrock through a VPC endpoint](../../assets/images/blog/amazon-bedrock-openai-models-codex-managed-agents-link.webp) _Figure: the network control in the article. Codex traffic can use PrivateLink so it does not cross the public internet to the model provider. Guardrails and encryption are named in the same control list and are not extra hops here. [Open the draw.io source](https://www.joinhorizon.co/examples/architecture-blog-2026/amazon-bedrock-openai-models-codex-managed-agents/link.drawio)._ The same controls that have made Bedrock the default GenAI platform for regulated industries apply to OpenAI models, Codex, and Managed Agents the moment they enter the platform: - **IAM and IAM Identity Center** for least-privilege access, including for individual developers using Codex and individual agents acting on behalf of services - **AWS PrivateLink** for VPC-private connectivity, so model traffic never traverses the public internet - **Bedrock Guardrails** for topic denial, PII detection and redaction, harmful content filters, and grounding checks, composable in front of OpenAI calls the same way they compose with Claude - **Encryption** at rest (KMS) and in transit (TLS) - **CloudTrail** logging for every API call, every Codex action, and every Managed Agent step - **Per-agent identities and per-action logs**, which AWS calls out specifically for Managed Agents, so an agent doing autonomous work has the same observability as a human user For teams that have already invested in [Bedrock Guardrails](https://www.joinhorizon.co/blog/how-to-set-up-amazon-bedrock-guardrails-production/) and a VPC-private architecture, OpenAI workloads slot into the same controls without re-architecting. ## How this reshapes the "Bedrock vs. OpenAI API" decision We published [AWS Bedrock vs. OpenAI API: Enterprise Decision Guide 2026](https://www.joinhorizon.co/blog/aws-bedrock-vs-openai-api-enterprise/) earlier this month. The framing was: pick OpenAI direct for raw capability and OpenAI familiarity, pick Bedrock for compliance, model choice, and AWS-native integration. That trade-off has narrowed. For a regulated enterprise the current framing is: | If you want… | Pre-announcement | Post-announcement | | -------------------------------------------- | --------------------------- | -------------------------------- | | OpenAI's frontier models | OpenAI direct API | **Bedrock (preview)** | | HIPAA-eligible OpenAI usage | Not viable | **Bedrock** | | OpenAI spend on AWS commitments | Not possible | **Bedrock** | | Codex with corporate SSO | Personal keys + workarounds | **Bedrock** | | Managed agent runtime on OpenAI | DIY on EC2 / ECS | **Bedrock Managed Agents** | | Lowest-latency raw OpenAI calls outside AWS | OpenAI direct API | OpenAI direct API | | Earliest access to brand-new OpenAI features | OpenAI direct API | OpenAI direct API (then Bedrock) | OpenAI direct still wins for two profiles: teams whose primary workload is outside AWS (Azure-first, GCP-first, or on-prem), and teams that need bleeding-edge model features the day they ship. For everyone else with a meaningful AWS footprint, Bedrock is now the default path _to_ OpenAI rather than a substitute _for_ OpenAI. ## Cost and commitment math Economics drive most of the boardroom decisions here. - **Spend consolidation.** OpenAI usage on Bedrock is consumed under your AWS account and applied toward your existing AWS cloud commitments. If you are already on an Enterprise Discount Program or a private pricing agreement, OpenAI spend now counts toward that commitment instead of competing with it. - **One invoice, one procurement.** Vendor management drops from "OpenAI + AWS" to "AWS," and legal, finance, and security reviews compress accordingly. - **Savings Plan adjacency.** While AWS has not announced a Savings Plan specifically for OpenAI tokens, Bedrock has historically introduced provisioned throughput / committed-use options for high-volume models. We expect a similar mechanism over time. - **Hidden cost: preview pricing.** When this post first shipped, pricing was not published. OpenAI model rates are now on the [Bedrock pricing](https://aws.amazon.com/bedrock/pricing/) page; confirm Managed Agents preview pricing there before you commit to a spend forecast. For broader cost discipline on Bedrock workloads, see our [token budgets and model selection playbook](https://www.joinhorizon.co/blog/aws-bedrock-cost-optimization-token-budgets-model-selection/). ## What to do this quarter Three tracks, in priority order: **1. Register for the Limited Preview.** AWS gates preview access by account standing, region, and workload fit. Submit a request with a real workload description (token volume, region, security posture) rather than a generic "we want access." FactualMinds runs this kind of preparation for clients in week one. **2. Audit your existing OpenAI usage.** Inventory every place an `OPENAI_API_KEY` lives in your code, your CI, your developer laptops, and your SaaS integrations. You will need this map the moment Bedrock OpenAI access is granted, and you will need it sooner regardless, because the moment your CFO sees that Bedrock OpenAI counts toward AWS commitments, the consolidation conversation starts. **3. Plan an agent pilot on AgentCore.** Pick a single high-value, well-bounded workflow (a coding agent on a specific repo, a procurement agent on a specific catalog, a triage agent on a specific queue). Scope a 4 to 6 week pilot once preview access lands. Treat it as the test fixture for whether Managed Agents replace your DIY agent code, augment it, or coexist. ## When to wait, when to pilot Pilot now if you are an AWS-first enterprise, you have a regulated workload that has been waiting for an OpenAI option, your developers are already running Codex with personal keys, or you have a long-running agent workload where AWS owning the runtime would meaningfully reduce ops burden. Wait if your primary cloud is Azure or GCP, your workload requires the absolute latest OpenAI feature on day one, you have no AWS commitment to amortize against, or you have not yet adopted Bedrock at all (start with [Knowledge Bases](https://www.joinhorizon.co/blog/how-to-build-rag-pipeline-amazon-bedrock-knowledge-bases/) and [Guardrails](https://www.joinhorizon.co/blog/how-to-set-up-amazon-bedrock-guardrails-production/) before reaching for managed agents). A few things were still unknown as of late April 2026: which exact OpenAI model SKUs will be exposed, which AWS regions will host them at GA, the published on-demand token pricing, the SLAs around Managed Agents, and the AgentCore roadmap for authorization policies and agent discovery. Architect with abstractions (model routing layers, IAM-bounded agent identities) so you can absorb those answers as they land. ## How FactualMinds helps We are an AWS Select Tier Services Partner running production Bedrock deployments for healthcare, SaaS, and ecommerce clients. For the OpenAI-on-Bedrock launch specifically, we help teams with three things: - **Preview access strategy:** preparing the request, the regions, the workload narrative, and the security posture AWS wants to see - **Migration from OpenAI direct:** auditing existing OpenAI usage, mapping it to Bedrock equivalents, redirecting traffic without breaking applications - **Managed Agent pilots:** scoping the first AgentCore-hosted agent, wiring IAM and Guardrails, instrumenting CloudTrail, and shipping in 4 to 6 weeks If you are evaluating this announcement for your stack, [see our AWS Bedrock consulting services](https://www.joinhorizon.co/services/aws-bedrock/) or [talk to our team](https://www.joinhorizon.co/contact-us/). ## Frequently asked questions The FAQ block above answers the most common questions we are hearing: - Which OpenAI models are available on Amazon Bedrock? - How is Codex on Bedrock different from using Codex directly? - What are Bedrock Managed Agents powered by OpenAI? - Does OpenAI usage on Bedrock count toward AWS commitments? - How do Bedrock Managed Agents handle security and auditability? - How can my team get access to the Limited Preview? For deeper architecture questions (model selection, agent topology, cost modeling, or migration sequencing), [reach out](https://www.joinhorizon.co/contact-us/) and we will scope a working session with the FactualMinds Bedrock team. ## Related reading - [How to Set Up Amazon Q for Business with SharePoint and S3](https://www.joinhorizon.co/blog/how-to-set-up-amazon-q-for-business-sharepoint-s3/) ## FAQ ### Which OpenAI models are available on Amazon Bedrock? AWS has confirmed that OpenAI's latest frontier models are accessible through Amazon Bedrock under a Limited Preview announced on April 28, 2026. The announcement references 'the latest OpenAI models' and 'OpenAI frontier models' generically rather than naming specific SKUs. The Codex coding agent, used by more than four million people each week, is included via the Bedrock API and works through the Codex CLI, desktop app, and Visual Studio Code extension. Exact model names, regions, context windows, and on-demand pricing are expected to firm up as the preview moves toward general availability. Until then, treat capacity as preview-tier and avoid hard-coding model IDs into production paths. ### How is Codex on Bedrock different from using Codex directly? Codex on Bedrock authenticates with AWS credentials instead of personal OpenAI API keys. That single change has several enterprise effects: developers sign in through your existing AWS SSO/IAM Identity Center, every Codex action is attributable to a real human and logged in CloudTrail, network traffic can be pinned to AWS PrivateLink, and Codex spend rolls into the AWS bill and into your existing AWS commitments. Functionally the developer experience is the same (CLI, desktop, VS Code extension), but operationally it stops being a shadow-IT problem and becomes a governed AWS service. ### What are Bedrock Managed Agents powered by OpenAI? Bedrock Managed Agents are production-grade AI agents built on OpenAI's frontier models and agent framework, hosted by AWS on Amazon Bedrock AgentCore. Each agent has persistent memory across sessions, skills that encode procedures, an IAM identity that enforces permissions, and AgentCore compute that AWS scales for you. Update (September 29, 2026): the public preview docs show a different split than this April description. AWS runs the conversation and model calls, but commands and tools run on compute you provide (your own host running codex exec-server, or an AgentCore Runtime in your account), and the preview has no built-in long-term memory. See our Bedrock Managed Agents preview guide for the details. They are the right starting point when you want OpenAI capability with AWS governance and no MLOps. ### Does OpenAI usage on Bedrock count toward AWS commitments? Yes. AWS has explicitly stated that customers can apply OpenAI model usage on Bedrock toward their existing AWS cloud commitments, including Enterprise Discount Programs, private pricing agreements, Savings Plans context, and consolidated billing. For enterprises that already have a multi-million-dollar AWS commitment, this is the main change: OpenAI consumption stops being a separate procurement line item and becomes part of AWS spend you have already committed. It also collapses the vendor management overhead from two billing relationships to one. ### How do Bedrock Managed Agents handle security and auditability? Each Managed Agent runs with its own IAM identity and logs every action it takes. The agent runtime sits inside Amazon Bedrock with the same enterprise controls customers already use for Claude, Llama, and Nova models on the platform: IAM-based access, AWS PrivateLink for private connectivity, Bedrock Guardrails for content safety and PII redaction, encryption at rest and in transit, and CloudTrail for full audit logs. AgentCore is also expected to add authorization policies, agent discovery, and observability tooling as the platform matures. ### How can my team get access to the Limited Preview? AWS has opened a registration form on the Bedrock product page for OpenAI models, Codex, and Managed Agents. Approval is gated by AWS account standing, region availability, and use-case fit. If you are an AWS customer with a meaningful commitment and a concrete AI workload, you are well positioned. FactualMinds, as an AWS Select Tier Services Partner, helps clients prepare a request that reads well to AWS (workload description, expected token volume, regions, security posture) and then run a 4 to 6 week pilot once access lands. --- *Source: https://www.joinhorizon.co/blog/amazon-bedrock-openai-models-codex-managed-agents/*