A vision-language model (VLM) is a model that accepts images alongside text and produces text responses that reason about the visual content. All major frontier models now have this capability: GPT-4o, Claude 3+ series, Gemini 2.5, and Llama 3.2 Vision.
Production use cases that have matured: reading scanned PDFs and invoices (extract structured data from unstructured documents), screenshot-based UI testing (describe what's wrong in a UI), product image tagging (classify and describe product photos at scale), form extraction (read handwritten or printed forms), and accessibility (generate alt text for images).
Important to know about costs: a high-resolution image can consume 1,000-2,000+ tokens depending on how the model encodes it. Resize and crop images to the minimum necessary resolution before sending — for text-heavy documents you rarely need more than 1024px on the longest side.
Bring this to your business
Knowing the term is one thing. Shipping it is another.
We do two-week AI Sprints — one term, one workflow, into production by Day 10.