Hi, I’m building an applied ai application that takes in a series of images as input and outputs a json object that adheres to a reponse_format pydantic model. I’m using the beta.chat.completions.parse method and it’s working great for some smaller docs (<30 images) but stops returning a response after that. The input tokens don’t reach the context window limit (75k) and no debuggable error is being returned. Anyone know why this is happening?
Seperately, noticing in the image analysis docs a mention to not send images to the model that contain text. Why is this? TIA https://platform.openai.com/docs/guides/images#image-input-requirements