Amazon Web Services (AWS) has announced the integration of OpenAI’s open-weight models into its AI platforms, Amazon Bedrock and Amazon SageMaker. This marks the first time OpenAI’s models are being offered on AWS, expanding the reach of powerful generative AI capabilities to a broader range of enterprise users and developers. The models now available—gpt-oss-20b and gpt-oss-120b—have been designed for high performance across tasks such as coding, math, reasoning, and agentic workflows, enabling more sophisticated and intelligent applications across industries.
This comes just a day after OpenAI announced the availability of gpt-oss for everyone to download and use. These models are optimized not just for performance but also for efficiency, making them particularly appealing to enterprises looking to deploy AI tools at scale. AWS has emphasized the models’ competitive edge by stating they outperform other leading offerings in terms of price-to-performance, including Meta’s Llama 3, DeepSeek-R1, and even OpenAI’s own o4 model.
By leveraging AWS’s infrastructure, customers can now use these models with the reliability, scalability, and ease-of-use that cloud-based services offer, without the challenges of managing complex AI deployments in-house.
The models are accessible through Amazon Bedrock, which allows developers to build and scale generative AI applications using foundation models via an API, and Amazon SageMaker, a platform for building, training, and deploying machine learning models.
AWS has also integrated features such as Guardrails, which help monitor and manage the safety of AI outputs, ensuring enterprises can deploy solutions that are not only powerful but also responsible. Additional enterprise tools like Knowledge Bases and Custom Model Import are expected to further enhance the flexibility and control available to customers.
