OpenAI Vision model

GPT-4 with Vision (GPT-4V) is the multimodal model that processes and interprets image inputs: it sees, understands, and reasons over complex visual data.

The OpenAI Vision model, specifically GPT-4 with Vision (GPT-4V), is a powerful multimodal system that integrates language and advanced computer vision. It processes image inputs (via URL or Base64) alongside text prompts, enabling complex visual reasoning and analysis. For example, the model can analyze a photograph of a traffic pole festooned with signs and correctly determine a complex parking rule, such as 'one hour starting at 4PM'. This capability extends to object recognition, scene description, and optical character recognition (OCR), allowing developers to build sophisticated applications like accessibility tools or an agent that turns a webpage screenshot into code.

https://platform.openai.com/docs/guides/vision