Models / Meta
Vision

NIM Llama 3.2 90B Vision Instruct

NVIDIA NIM for GPU accelerated Llama 3.2 90B Vision Instruct inference through OpenAI compatible APIs.

About model

NIM Llama 3.2 90B Vision Instruct processes visual and text-based inputs to generate human-like responses. It is the largest Llama vision model, ideal for applications requiring advanced vision understanding. Suitable for developers and researchers needing high-performance vision capabilities.

To run this model, you first need to deploy it on a Dedicated Endpoint.

    Related models
    • Model provider
      Meta
    • Type
      Vision
    • Main use cases
      Vision
    • Deployment
      On-Demand Dedicated
      Monthly Reserved
    • Parameters
      90B
    • Context length
      128K
    • Input modalities
      Text
      Image
    • Output modalities
      Text
    • Released
      September 24, 2024
    • Last updated
      August 26, 2025
    • External link
    • Category
      Vision