We're sharing new research on how models hack public benchmarks. The latest models, including Opus 4.8 and Composer 2.5, learn to retrieve solutions from the internet or git history. When we apply a stricter harness, eval scores drop significantly. Read more here: https://lnkd.in/giFNCMCV
About us
Cursor is a coding agent for building ambitious software. Our goal is to help you engineer anything. Our work includes training the world’s most widely used coding models, creating infrastructure that supports billions of requests per day, and building better ways for humans and AIs to work together.
- Website
-
http://cursor.com
External link for Cursor
- Industry
- Software Development
- Company size
- 501-1,000 employees
- Type
- Privately Held
Employees at Cursor
Updates
-
You can now delegate tasks to Cursor directly from Notion. It's built on the Cursor SDK, so every cloud agent runs on the same models, harness, and runtime that power Cursor. @Cursor on any spec or assign it a task to open a PR your whole team can review. More on how Notion built it with the Cursor SDK: cursor.com/blog/notion
-
Plugins, skills, and MCPs help make Cursor powerful and customized for your team. We’ve added the Customize page to bring them into one place. You can now: • See what your team uses most on a leaderboard, then add any plugin, skill, or MCP server in one click. • Have Cursor build interactive dashboards and visualizations from partner plugin data in Atlassian, Hex, and more. • Manage plugins, skills, MCP servers, subagents, rules, commands, and hooks at the user, team, or workspace level. Learn more here: https://lnkd.in/g3t2FyED
-
75% of pull requests at Coinbase are now created by agents. Over 2,400 developers at Coinbase use Cursor as part of an agent-first engineering model. Since the start of the year, the average engineer is merging 55% more PRs and saving 7 hours of manual coding each week, with teams of 1-2 engineers now building features that once required a full team. "The product has become a mission control for agents rather than just a raw IDE," said Chintan Turakhia, Senior Director of Engineering at Coinbase. Read more: https://lnkd.in/gQcTMNyV
-
Three announcements from our keynote at Compile, including how we're training a new model with SpaceX. Watch it here: https://lnkd.in/geQvZDCE
-
-
Cursor Automations save you time by automating repetitive tasks with always-on agents. You can now create them directly from your local agent session with /automate. Describe the task you want to automate in plain language and Cursor will configure the triggers, instructions, and tools for you. This release also introduces new triggers for GitHub and Slack, and support for computer use. Learn more here: https://lnkd.in/g7-RmF73
-
It’s now easier to move local agents to the cloud so they can keep working with your laptop closed. Prompt Cursor from your phone, run many agents in parallel, and get back PRs with demos of their work. See everything new here: https://lnkd.in/g2nYUCqb
-
We're launching code storage and git hosting. Origin gives teams and agents a place to host, review, and collaborate on code. Available this fall. Join the waitlist. cursor.com/origin
-
Wayfair is using Cursor to compress months of ML research into days. In December 2025, five researchers tested 110 variants of a tag-validation model across a four-day sprint and cut inference costs by 94%. Three months later, they ran the same playbook with newer models and cut costs by another 90%. "Cursor changed the bottleneck from 'How long will this take to build?' to 'What is the next idea worth testing?' That is a much better place for a scientist to spend their attention," said Omer L., Senior Machine Learning Scientist at Wayfair. Learn more here: https://lnkd.in/es3m-rPf
-
Auto-review is now the default for all new users. A classifier subagent reviews actions in context before deciding whether to allow, block, or ask for approval. Our evals show it's 97% accurate, with most misses near ambiguous edges. More on how we built it: https://lnkd.in/gaHd79wB