Oracle Cloud Infrastructure Documentation

Supported Google Models

You can import large language models from Hugging Face and OCI Object Storage buckets into OCI Generative AI, create endpoints for those models, and use them in the Generative AI service.

Google Gemma series, including Gemma2 and Gemma3 open source models are built for broad language processing needs and high versatility. For more information, see Gemma2 and Gemma3 in the Hugging Face documentation.

Gemma

Supported Gemma Models
Hugging Face Model ID Model Capability Recommended Dedicated AI Cluster Unit Shape
google/gemma-3-4b-it IMAGE_TEXT_TO_TEXT A100_80G_X1
google/gemma-3-270m-it TEXT_TO_TEXT A100_80G_X1
google/gemma-3-27b-it IMAGE_TEXT_TO_TEXT A100_80G_X2
google/gemma-3-12b-it IMAGE_TEXT_TO_TEXT A100_80G_X1
google/gemma-3-1b-it TEXT_TO_TEXT A100_80G_X1
google/gemma-2-9b-it TEXT_TO_TEXT A100_80G_X1
google/gemma-2-27b-it TEXT_TO_TEXT A100_80G_X2
google/gemma-2-2b-it TEXT_TO_TEXT A100_80G_X1
Important

  • While you can import any chat, embedding, (and fine-tuned) model supported through Open Model Engine (with vLLM or SGLang runtime), only models explicitly listed on this page are supported for this model family. Unlisted models might have compatibility issues and we recommend that you test any unlisted model before production use. Learn about OCI Generative AI Imported Model Architecture.

  • Imported models support the native context length specified by the model provider. However, the effective maximum context length is also limited by OCI Generative AI's underlying hardware setup. To take full advantage of a model's native context length, you might need to provision more hardware resources.
  • Fine-tuned models are supported only if they match the supported base model's transformer version and have a parameter count within ±10% of the original.
  • For available hardware and steps on how to deploy the imported models, see Managing Imported Models.
  • If the recommended unit shape isn't available in the region, select a higher-tier option. For example, if A100 isn't available, select H100.