Command A Vision

Parameters

112B

Published

Jul 31, 2025

Command A Vision is an AI model developed by Cohere (Canada), first published in July 2025. It works in the vision, multimodal and language domain, on tasks such as visual question answering, character recognition (ocr), language modeling/generation and question answering.

Epoch AI has no training-compute estimate for this model. The model has 112,000,000,000 parameters.

Access: Open weights (non-commercial). Its weights are openly available. It is built on top of Cohere Command A,SigLIP 2. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Cohere
Country of organization: Canada
Domain: Vision, Multimodal, Language
Task: Visual question answering, Character recognition (OCR), Language modeling/generation, Question answering
Parameters: 112,000,000,000
Model accessibility: Open weights (non-commercial)
Open weights: Yes
Base model: Cohere Command A, SigLIP 2
Epoch confidence: Confident