Summary:
Cursor launches long-running agents in preview
Why It Matters:
Cursor is moving beyond simple code edits toward autonomous agents that can handle complex, multi-step tasks over longer periods. This signals a shift from AI as a "copilot" to AI as an "independent worker."
What Changed:
- Added a new changelog entry for "Long-running Agents in Research Preview" (Feb 12, 2026).
- Updated the homepage title tag to "Cursor: The best way to code with AI."
- Updated the model selection demo to show "GPT-5.2" as the active model.
- Modified the interactive demo text to show more advanced search queries like "Where is the model picker UI implemented?"