Per the docs, Opus 4.5 supports a new effort param:
What effort is opus-4.5-thinking using? (in cursor agent)
1 Like
Claude Opus 4.5 uses the effort parameter toscale its reasoning budget. Increasing effort expands the model’s internal deliberation steps, improving reasoning depth and accuracy at the cost of longer generation time.
Is there any plan to introduce new variants of Opus 4.5 with medium and low effort into cursor?