Llama

Compute Owner ENTITY AGENT
1

Meta Llama

Recommended llama models:

2

Llama 3.2 Vision 11B

As of 2025-03 the llama 3.2 vision 11b is one of the standard model available in team compute deployments worldwide.

Features:

  • understands input of text and images
  • context length of 128K tokens
  • multiple languages

Note to use the 128K context length you need to override Ollama's default 2K length.

Vision

The maximum image size is 1120x1120 pixels, with a 2048 token output limit and 128k context length.