Amazon Web Services (AWS) has just announced a significant expansion to its AI offerings: OpenAI’s open weight foundation models—gpt-oss-120b and gpt-oss-20b—are now available on Amazon Bedrock and Amazon SageMaker JumpStart. This strategic move is set to revolutionize how organizations of all sizes build and scale generative AI applications, by offering unprecedented access to powerful AI tools in a fully managed and secure environment.
What Are Open Weight Models—and Why They Matter
Unlike traditional closed models, open weight models give developers full access to the neural network weights, enabling greater customization, control, and transparency. This is particularly important for industries needing tailor-made AI solutions or compliance with data governance policies.
Now, with gpt-oss-120b and gpt-oss-20b natively integrated into AWS platforms, companies can take advantage of OpenAI’s advanced reasoning capabilities, while benefiting from the scalability, performance, and enterprise security of the AWS ecosystem.
Performance that Outshines the Competition
The gpt-oss-120b model stands out as a leader in cost-efficiency. According to AWS, it’s:
-
3x more price-performant than Google’s Gemini model
-
5x more efficient than DeepSeek-R1
-
2x more performant than OpenAI’s own o4 model
This makes OpenAI’s new open models a strong choice for developers and organizations seeking high-performance AI at scale—without breaking the bank.
Powerful AI Agents with Bedrock and AgentCore
The real magic begins when combining these models with Amazon Bedrock AgentCore, a toolkit that allows customers to build, deploy, and scale AI-powered agents across use cases like customer service, workflow automation, and data analysis.
With AgentCore and OpenAI’s open models, businesses can now create highly intelligent, responsive agents—all while operating within the robust security infrastructure of AWS.
Enterprise Features and Guardrails for Responsible AI
Security and safety remain central to the offering. AWS enables seamless integration of these models with enterprise-grade safeguards, including:
-
Guardrails that block up to 88% of harmful content
-
Custom Model Import for unique enterprise data
-
Knowledge Bases for retrieval-augmented generation
-
Fine-tuning and deployment tools via SageMaker
These features make it easier for businesses to deploy AI responsibly while maintaining compliance and protecting user data.
Strategic Partnership: OpenAI x AWS
This collaboration marks OpenAI’s first entry into AWS’s Bedrock and SageMaker ecosystems. According to Atul Deo, director of product at AWS:
“The addition of OpenAI as our newest open weight model provider marks a natural progression in our commitment to bringing cutting-edge AI to organizations worldwide.”
Echoing this sentiment, Dmitry Pimenov, product lead at OpenAI, added:
“Our open weight models help developers—from solo builders to large enterprise teams—unlock new possibilities across industries and use cases.”
Why It Matters
With this integration, millions of AWS customers now gain access to powerful, flexible AI models backed by OpenAI’s leading technology. This move not only accelerates innovation but democratizes access to cutting-edge tools, setting a new standard in the evolving generative AI landscape.
