I tried advanced o1 models (with usage-based pricing for o1-preview) and have two questions relating using these models in Cursor environment.
About my setting: Winfows 11 Home, Cursor version 0.42.4.
1). Can I call o1 models in Cursor chat the same way as I call them in chatgpt.com (Pro subscription is needed) site? So, I mean, I want to have option to cancel any wrappers, any additional prompts from Cursor. Can I do this “raw call” in Cursor chat?
This would allow to refuse direct subscription to OpenAI. But now, it seems to me that sometimes the quality of work of o1 models in Cursor deteriorates somewhat.
2). I have a problem to apply suggestions (to fix a code) of o1 models. When I click on the Apply button, it tries to create a new file and paste the corrected code into it. Apply button works well after Claude Sonnet (and other models) suggestions, but fails after o1 models. What can be done?