What is it?
Our new GPT-4 Turbo model now supports image-to-text. We offer two pricing options to choose from on a per-image basis, which depend on the input image size.
How do I access it?
GPT-4 with vision is currently available to all developers who have access to GPT-4. The model name is gpt-4-vision-preview via the ChatCompletions API. For further details on how to calculate cost and format inputs, check out our vision guide.
What are the rate limits for this model? Can I increase them?
Rate limits on the gpt-4-vision-preview model will be fixed to 20 requests per minute and 100 requests per day. Your tokens per minute limits are identical to your gpt-4 limits. It’s not possible to increase rate limits on the preview model. A stable model will become available in the coming weeks.