Lanxk's Blog, sharing and recording AI programming
Cursor's experience with the Claude 4 sonnet thinking model
1 min
A few days ago, when Claude 4 dropped, Cursor hopped on it right away, and after using it for two days, it’s pretty decent!Compared to the claude-3.7-sonnet(thinking) model, here are the immediate pros I’ve noticed (still updating):
It gets dev needs better—no more random or repetitive tweaks like 3.7, fewer unnecessary changes, so it feels faster than 3.7;
It can think ahead—spot a bug on a page (like missing null checks) and fix similar issues across the board;
Less likely to mess up—used it to optimize a performance issue, and the tweaks were spot-on, barely needed any follow-ups;
It auto-generates a PERFORMANCE_OPTIMIZATION.md doc (though it misplaced it) detailing the optimization suggestions, process, code changes, and next steps.
Anthropic released Claude Opus 4.1 on August 5th (Beijing time), and by 3 AM on August 6th, Cursor already supports it with the option enabled by default in settings
Cursor 1.5.5 version download links and Cursor v1.5 changelog: Added a Toggle AI Pane icon on the right side of the top menu bar, startup page now displays Chat sidebar, Cursor settings page's 'Tools & Integrations' changed to 'MCP & Integrations'
OpenAI released GPT-5 on Aug 7. Cursor added support immediately, with eight variants: GPT-5, GPT-5-fast, GPT-5-high, GPT-5-low, GPT-5-high-fast, GPT-5-low-fast, GPT-5-nano, and GPT-5-mini.
Today I discovered that Cursor has automatically updated to v1.4.0, and just now I received another update notification. After restarting, I found it had been updated to v1.4.1.