Skip to content
KoishiAI
ไทย
← Back to all articles
aws openai amazon-bedrock gpt-5 ai-agents

AWS Launches GPT-5.5 and OpenAI Frontier on Bedrock Following $50B Deal

AWS is now offering GPT-5.5, GPT-5.4, and the OpenAI Frontier agent platform on Amazon Bedrock, marking the first time OpenAI's frontier models are available outside of Microsoft Azure.

KoishiAI · Reviewed by: เกียรติดำรง ตรีครุธพันธ์ · · 4 min read
AI-drafted from cited sources, fact-checked and reviewed by a human editor. How we work · Standards · Report an error

TL;DR: AWS is making OpenAI’s GPT-5.5 and GPT-5.4 models available on Amazon Bedrock, alongside the new OpenAI Frontier agent platform. This launch follows a $50 billion investment deal between Amazon and OpenAI, ending OpenAI’s previous exclusivity with Microsoft Azure. The partnership includes a co-created Stateful Runtime Environment and a commitment from OpenAI to consume 2 gigawatts of AWS Trainium.

Key facts

  • AWS launched access to OpenAI’s GPT-5.5 and GPT-5.4 models via Amazon Bedrock in a limited preview on April 28, 2026 [3][6][7].
  • AWS is the exclusive third-party cloud distribution provider for OpenAI Frontier, a platform for building and managing teams of AI agents [5][8].
  • Amazon and OpenAI finalized a $50 billion multi-year investment and partnership agreement in late February 2026 [2][5][8].
  • OpenAI plans to consume 2 gigawatts of AWS Trainium capacity to support the advanced workloads of the partnership [2][5][8].
  • AWS and OpenAI are co-creating a Stateful Runtime Environment on Amazon Bedrock to enable persistent memory and identity for AI agents [5][8].
  • OpenAI’s coding agent, Codex, and the new Bedrock Managed Agents service are now available to enterprise customers [6][7].
  • The partnership includes the development of customized models to power Amazon’s customer-facing applications [5][8].

AWS Expands OpenAI Access with GPT-5.5 and Frontier Platform

Amazon Web Services (AWS) has significantly expanded its generative AI portfolio by launching access to OpenAI’s latest frontier models, including GPT-5.5 and GPT-5.4, on Amazon Bedrock [3][6][7]. This move, announced in a limited preview starting April 28, 2026, marks a pivotal shift in the cloud AI landscape, as it represents the first time OpenAI’s most advanced models are available outside of its long-standing exclusive partnership with Microsoft Azure [7]. The launch is the immediate operational result of a $50 billion multi-year investment and strategic partnership that Amazon and OpenAI finalized in late February 2026 [2][5][8].

The End of Exclusivity

The availability of OpenAI models on AWS follows reports that OpenAI’s previous exclusivity agreement with Microsoft has been revised, allowing OpenAI to distribute its products through other cloud providers [7]. This strategic pivot positions AWS as the exclusive third-party cloud distribution provider for OpenAI Frontier, a new platform designed for building and managing teams of AI agents [5][8].

OpenAI Frontier is a key component of this partnership, offering a dedicated environment for enterprises to develop and deploy complex agent-based workflows. By making Frontier available on AWS, OpenAI is providing a robust infrastructure for the next generation of AI applications, which require more than just single-model inference [5][8].

New Models and Services on Bedrock

The latest offering on Amazon Bedrock includes OpenAI’s GPT-5.5 and GPT-5.4 models, which support a 128K context window and adjustable reasoning levels [1][3]. These models are now accessible to developers and enterprises through the Bedrock console, API, and SDKs [3]. Additionally, OpenAI’s coding agent, Codex, is available on Amazon Bedrock to support enterprise software development workflows [6][7].

A significant addition to the AWS-OpenAI partnership is the introduction of Bedrock Managed Agents, a new service powered by OpenAI that enables developers to build production-ready AI agents [3][6][7]. This service is designed to simplify the creation of autonomous agents that can perform complex tasks across multiple applications and data sources.

Furthermore, AWS and OpenAI are co-creating a Stateful Runtime Environment available on Amazon Bedrock [2][5][8]. This environment is crucial for building production-scale applications with persistent memory and identity, allowing AI agents to maintain context and state across multiple interactions and sessions [5][8].

Infrastructure and Custom Models

Under the terms of the deal, OpenAI is committed to consuming 2 gigawatts of AWS Trainium capacity to support these advanced workloads [2][5][8]. This infrastructure investment underscores the scale of the partnership and AWS’s role in supporting the computational demands of frontier AI models [2].

The partnership also involves the development of customized models to power Amazon’s customer-facing applications [5][8]. This collaboration will likely result in tailored AI solutions that leverage both OpenAI’s model capabilities and AWS’s cloud infrastructure to enhance Amazon’s own services.

Open Source Models

In addition to the frontier models, AWS has also made OpenAI’s open-weight models, specifically gpt-oss-120b and gpt-oss-20b, available on Amazon Bedrock and SageMaker JumpStart [1]. These models provide developers with flexible options for fine-tuning and deploying AI applications with greater control over their data and models [1].

Strategic Implications

The launch of OpenAI models on AWS signals a potential shift in the dominance of Microsoft Azure in the AI cloud market. By securing access to OpenAI’s latest models and the Frontier platform, AWS is strengthening its position as a leading provider of enterprise AI solutions [7]. This move is expected to accelerate the adoption of AI agents and advanced language models across industries, as developers gain access to a broader range of tools and infrastructure [3][6].

The partnership between AWS and OpenAI represents a significant milestone in the evolution of generative AI, highlighting the increasing importance of multi-cloud strategies and the collaborative development of AI infrastructure [5][8]. As enterprises continue to explore the potential of AI agents and advanced language models, the availability of these tools on AWS will play a crucial role in shaping the future of the industry [3][7].

Sources

  1. OpenAI frontier models on Amazon Bedrock – AWS (aws.amazon.com) — 2026-04-28
  2. OpenAI Models on Amazon Bedrock: AWS expands partnership with Codex and Managed Agents (www.aboutamazon.com) — 2026-04-28
  3. Amazon is already offering new OpenAI products on AWS (tech.yahoo.com) — 2026-04-28
  4. Amazon And OpenAI Forge US$50b AI Deal Centered On AWS Growth (finance.yahoo.com) — 2026-02-28
  5. OpenAI and AWS Partnership for the new “OpenAI Frontier” (community.openai.com) — 2026-03-06
  6. Amazon, OpenAI announce strategic partnership (www.aboutamazon.com) — 2026-02-27
  7. OpenAI open weight models now available on AWS | Amazon Web Services (aws.amazon.com) — 2025-08-05

Frequently asked questions

What OpenAI models are now available on AWS? ▼
AWS is offering OpenAI's GPT-5.5 and GPT-5.4 models, as well as the open-weight models gpt-oss-120b and gpt-oss-20b, on Amazon Bedrock and SageMaker JumpStart [1][3][6]. Additionally, the OpenAI coding agent, Codex, is available for enterprise development [6].
When did AWS and OpenAI announce their partnership? ▼
Amazon and OpenAI announced their $50 billion multi-year investment and strategic partnership in late February 2026 [2][5][8]. The availability of the new models on AWS began in a limited preview on April 28, 2026 [3][6][7].
What is OpenAI Frontier and is it available on AWS? ▼
OpenAI Frontier is a platform for building and managing teams of AI agents [5]. AWS is the exclusive third-party cloud distribution provider for OpenAI Frontier, making it available to developers on the AWS platform [5][8].
How does AWS support OpenAI's infrastructure needs? ▼
OpenAI is committed to consuming 2 gigawatts of AWS Trainium capacity to support the advanced workloads of the partnership [2][5][8]. This infrastructure supports the computational demands of frontier AI models and agent-based applications [2].
What is the Stateful Runtime Environment? ▼
The Stateful Runtime Environment is a co-created feature on Amazon Bedrock that enables developers to build production-scale applications with persistent memory and identity for AI agents [5][8]. It allows agents to maintain context and state across multiple interactions [5].