GPT vision api is behind. We should switch to Claude. Also base64 encoding an image and uploading fails on larger images. We should switch to multi-part upload
GPT vision api is behind. We should switch to Claude.
Also base64 encoding an image and uploading fails on larger images. We should switch to multi-part upload