Cursor

Our take

Cursor set the template that every AI editor is now chasing: take VS Code, strip it down, rebuild the AI layer from scratch, and ship fast enough that the competition can't catch up. Three years in, that lead is holding, though the landscape looks different than it did at launch.

The defining shift was Cursor 2.0 (October 2025), which redesigned the interface around managing coding agents rather than editing individual files. You can run several agents in parallel across git worktrees, remote machines, or cloud VMs, kick one off on a feature branch, and come back to a merge-ready pull request. It's a real step past "autocomplete with extra steps." Underneath it, Cursor's in-house Composer model has advanced to Composer 2.5, which delivers frontier-level agent performance and, importantly, runs unlimited through Auto mode without spending credits.

That credit model is the main point of friction. Every paid plan now includes a credit pool equal to its price: Pro is $20/month with a $20 pool, Pro+ is $60, and Ultra is $200. Auto mode is unlimited and free, but the moment you hand-pick a frontier model like Claude Sonnet or GPT, you draw from that pool at API rates. For daily agent users that math tilts toward Pro+ or Ultra faster than the $20 headline suggests, so treat Pro as a starting point rather than the ceiling for heavy AI work.

For developers already comfortable in VS Code, the switching cost is effectively zero. Your extensions, keybindings, and themes carry over unchanged. That's a genuine moat. The question isn't whether Cursor is a good editor; it's whether the usage math works out for your particular workflow.