Console

Llama 3.2 90B

A medium-sized 90B multimodal model that can support image reasoning, such as chart and graph analysis as well as image captioning.

Supported regions
Model ID	`llama-3.2-90b-vision-instruct-maas`
Launch stage	Preview
Supported inputs & outputs	Inputs: Text, Code Outputs: Text
Capabilities	Supported Llama Guard Not supported Batch predictions
Usage types	Supported Fixed quota Not supported Provisioned Throughput
Knowledge cutoff date	December 2023
Versions	`llama-3.2-90b-vision-instruct-maas` Launch stage: Preview Release date: September 25, 2024
	Model availability	United States `us-central1`
	ML processing	United States `Multi-region`
Quota limits	us-central1: QPM: 30 Max output: 8,192 Context length: 128,000
Pricing	See Pricing.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025-10-09 UTC.

Llama 3.2 90B Stay organized with collections Save and categorize content based on your preferences.