Import NVIDIA Nemotron 3 Ultra in OCI Generative AI

The NVIDIA Nemotron 3 Ultra model is now compatible for import in OCI Generative AI, following NVIDIA’s release of the model.

NVIDIA describes Nemotron 3 Ultra as a frontier-scale open model optimized for complex agentic workflows, long-context analysis, tool use, and high-accuracy reasoning across code, math, and science. The model has 550 billion total parameters with 55 billion active parameters and supports context lengths up to 1 million tokens.

In OCI Generative AI, use the following details for this model:

• Hugging Face model ID for import: nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4
• Recommended dedicated AI cluster unit shape: B200_X4

For the complete list of models compatible for import, see Compatible Models for Import. For available hardware unit sizes and deployment steps, see Managing Imported Models. For information about the service, see the Generative AI documentation.