All of the sudden Cursor agents are showing me all their internal planning thoughts…? Is this a recent change or did I flip a config item somehow?
Which models have you used before, which models are you using now? Each model has a different setup for Chain of Thought visibility, some even show it obsured (by showing just a summary written by another dedicated model).
Also, what version of Cursor are you on? Up to 1.1.4 I didn’t notice any changes.
I’m on 1.1.3 and it’s not letting me upgrade to 1.1.4 yet - it appears it was using claude-4-sonnet (via auto) which was doing the displayed reasoning stuff as of last night and not before. It was also wrecking my prototyping code so I’ve fixed it to 3.5 for now and will see what happens.
1 Like