OpenAI’s Open-Weight Models Now Available on AWS, Enhancing GenAI Capabilities

Prime Highlights

OpenAI open-weight models gpt-oss-120b and gpt-oss-20b are now supported on AWS via Amazon Bedrock and SageMaker JumpStart.
The models offer state-of-the-art reasoning, coding, and scientific capabilities with improved price-performance and compute-optimized efficiency.

Key Facts

For the first time, OpenAI models will be accessible beyond Microsoft Azure, increasing availability across cloud ecosystems.
The models offer up to 18x greater price-performance than some of the leading competition in generative AI.

Key Background

OpenAI has released two open-weight AI models, gpt-oss-120b and gpt-oss-20b, on Amazon Web Services (AWS). The models are available under Apache 2.0 license, which enables developers to fine-tune and use the models without usage restrictions. OpenAI is not releasing training data or code, but making model weights permissively licensed is a very big step towards making AI more transparent and accessible.

These models are integrated with Amazon Bedrock and SageMaker JumpStart, two of the strongest AWS platforms for hosting and running machine learning workloads. This enables developers to innovate, customize, and create generative AI models with AWS capabilities and infrastructure they are already familiar with. Amazon Bedrock offers easy deployment with standard APIs and safe environments, and SageMaker JumpStart includes pre-configured solutions for rapid deployment.

What differentiates these models is their state-of-the-art ability in reasoning, code generation, and scientific use cases. With a context window of as much as 128,000 tokens and an additional “Reasoning Mode” (tunable from low to med to high effort), the models can balance performance and latency depending on the workload. Chain-of-thought prompting and tool integration features also make them ideal for agent-based workloads and intricate workflows.

This announcement is significant in the sense that it brings to a halt OpenAI’s former monopoly on Microsoft Azure usage, a strategic move in leveraging cloud-based AI. Through its partnership with AWS, OpenAI is narrowing its scope and bringing its high-performance models closer. For AWS, this alignment tremendously strengthens its own position of market leadership in generative AI by having low-cost yet high-performance offerings against other large language models.

By this collaborative endeavor, AWS and OpenAI collectively are leading the development of enterprise-level generative AI.