The same question but get different response in cursor and chatgpt.
Does the 4o in cursor won’t use search ability? I think the chatgpt version of 4o use search automatically and get better response.
Yes, it’s different.
Cursor also has search features for the codebase and the web.
But if you’re using Cursor, it has a custom system message that makes your results different from using ChatGPT directly.
Also maybe token limits in cursor.
In the past, chatgpt.com with GPT-4 had way better results than over API.
ChatGPT uses a different models tuned for ChatGPT, it has also different integration, tools, access, has different prompts, restrictions etc…
Its not comparable using any model from API (like Cursor does) or from OpenAIs internal product ChatGPT.