Oracle Cloud Infrastructure (OCI) Generative AI is growing more versatile with two major updates: the general availability of OpenAI gpt-oss models as hosted options, and the release of the Model Import feature.
The new gpt-oss models are now generally available as hosted models within OCI Generative AI. gpt-oss models are leading open source GPT-style large language models (LLMs) that are fully managed by Oracle, giving you on-demand access with enterprise-grade reliability and a seamless, fully hosted experience. According to OpenAI, these outperform other open models of similar size on reasoning tasks, demonstrate advanced tool use capabilities, and are optimized for efficient deployment on consumer-grade hardware.
Alongside the expanding roster of hosted models, OCI Generative AI now provides you the option to bring your own LLMs with the new Model Import feature. With Model Import, you can import models directly from Hugging Face or from your OCI Object Storage into the service. This allows you to create endpoints for your chosen models and use them seamlessly within OCI, helping you to build and deploy AI solutions for your organization’s specific needs.
OCI Generative AI supports a wide range of chat model architectures for conversational AI, as well as embedding models that convert text and images into numerical vectors for advanced semantic understanding. [See the full list of supported model architectures here.] Model Import accelerates AI adoption by letting organizations deploy top open source and third-party language models within OCI’s secure, scalable environment.
These updates reinforce Oracle’s commitment to delivering an open, flexible, and secure ecosystem for enterprise AI innovation.
Want to know more?
For more details and pricing information, consult the Generative AI service documentation or contact your Oracle representative.
For more information, see the following resources: