Report
Llama 3.2 Vision Runs Locally in Ollama, Adding Image Analysis to Your Machine
Ollama now offers Llama 3.2 Vision in 11B and 90B sizes, letting you run image recognition tasks like handwriting OCR and chart reading on your own hardware.
Llama 3.2 Vision is now available to run in Ollama, in both 11B and 90B sizes. The announcement, dated November 6, 2024, means you can pull the model and start using it locally. To get started, download Ollama 0.4 and run `ollama run llama3.2-vision` for the 11B model, or `ollama run llama3.2-vision:90b` for the larger 90B version.
The model accepts images as input, enabling tasks such as handwriting optical character recognition (OCR), reading charts and tables, and answering questions about images. To add an image to a prompt, you can drag and drop it into the terminal, or on Linux, add a path to the image in the prompt.
Hardware requirements are significant: the 11B model needs at least 8GB of VRAM, while the 90B model requires at least 64GB of VRAM. This means the smaller model can run on many consumer GPUs, but the larger one demands a high-end workstation or server.
You can use Llama 3.2 Vision through the Ollama Python library, JavaScript library, or cURL. For example, a Python script can send an image and ask 'What is in this image?' The integration makes it straightforward to add image understanding to your own applications, all running locally on your machine.
Source details and supporting facts
Each line is stated by the page named above it.
Stated by ollama.com
- Llama 3.2 Vision 11B and 90B models are now available in Ollama.
- Llama 3.2 Vision 11B requires at least 8GB of VRAM, and the 90B model requires at least 64 GB of VRAM.
- Examples include handwriting OCR, charts & tables, and image Q&A.
- To add an image to the prompt, drag and drop it into the terminal, or add a path to the image to the prompt on Linux.
Sources
- Ollama blogText stored 15 September 2026
How this story was checked. Written from the 1 page listed above, stored 15 September 2026; claims checked against that stored text on 15 September 2026.
What that means
- 4 of 4 reported statements were confirmed against the page that carries them; the rest were removed rather than published.
- Figures in the text were required to appear in the stored source text: yes. Identifiers: yes.
- The check reads stored text only: no claim rests on a fresh look that did not happen.
- Where the reporting was silent, the text says so instead of filling the gap.