Cursor felt better at coding even with the same Claude models
In firsthand use of , Claude, and Cursor, Cursor felt like the most dependable tool for coding. was dropped after repeated frustrating experiences, even with version 2.0. After using free Claude for a long time, both Claude and Cursor paid plans were added.
With Sonnet 4.6 and 5, Cursor seemed sharper at writing code, more professional in analysis and , and more efficient than using Claude directly. No tasks, measurements, or detailed failures were provided, so this is a personal judgment rather than proof of a difference.
Key points
- was abandoned after frustrating experiences, including with version 2.0.
- The comparison came after long-term use of free Claude and to both Claude and Cursor.
- Cursor felt sharper for coding while using Sonnet 4.6 and 5.
- Cursor also seemed better at analysis, , and overall efficiency.
- There are no controlled tests or numbers showing that Cursor is objectively better.