Cursor

13 releases covered on ThursdAI · cursor.com ↗

May 2026

Cursor
New Models

Composer 2.5

Cursor launches Composer 2.5 with Opus-class coding at much lower cost

Cursor launched Composer 2.5, a coding model continued-trained on top of Kimi K2.5 (with permission) that delivers Opus-class coding performance at much lower cost. The crew noted Cursor is 'absolutely back' with strong pre-training and post-training teams, and that training now runs partly on the Colossus supercomputer.

April 2026

Cursor
Dev Tools

Cursor SDK

Cursor launches SDK exposing the runtime that powers the IDE

Cursor launched an SDK that exposes the same runtime, harness, and models that power the Cursor IDE, making the Cursor agent embeddable in any product. The Cursor Agent + GPT-5.5 combo also topped WolfBench's Terminal-Bench 2.0 leaderboard this week.

Cursor
Major Features & Updates

Cursor remote agents & code review agent

Cursor ships remote agents and a code review agent

Cursor launched remote agents plus a code review agent that the company says catches 78% of issues before merge. Mentioned in the week's tools and agentic-engineering roundup.

Cursor
Products & Apps

Cursor 3

Cursor 3 ships as agent-first rebuild, dropping the VS Code fork

Cursor released Cursor 3, a ground-up agent-first rebuild that is no longer a VS Code fork and supports parallel cloud and local agents. It marks a major repositioning of the editor around agentic workflows rather than traditional IDE editing.

March 2026

Cursor
New Models

Composer 2

Cursor Composer 2 beats Opus 4.6 on TerminalBench at a tenth of the price

Cursor launched Composer 2, its first proprietary model that genuinely competes with frontier labs. It scores 61% on TerminalBench (beating Opus 4.6) at $0.50/M input tokens, cheaper than GPT-5.4 Mini and 10x cheaper than Opus, running at 300+ tokens/sec. A fast variant costs 3x more for the same intelligence, kicking off a new 'fast mode' pricing trend where you pay a premium for speed rather than capability.

February 2026

Cursor
Major Features & Updates

Cloud Agents

Cursor launches cloud agents

Cursor launched cloud agents, moving agentic coding work off the local machine into remote, async sessions. The panel highlighted Cursor's cloud agents and UI demos as important progress for frontend development workflows.

December 2025

Cursor
Products & Apps

Cursor 2 + Composer

Cursor 2 and the Composer model level up IDE agents

Cursor shipped Cursor 2 along with its Composer model in October, leveling up in-IDE agentic coding. It capped a year in which Cursor's sales exploded on the back of Claude 3.7 and the vibe coding wave.

November 2025

Cursor
Major Features & Updates

Cursor in-IDE browser

Cursor adds a built-in browser inside the IDE

Cursor added an in-IDE browser, letting developers preview and interact with their running app without leaving the editor. The panel called out how performant the implementation is, tightening the loop between agentic code edits and visual verification.

October 2025

Cursor
Products & Apps

Cursor 2.0 & Composer

Cursor 2.0 ships with Composer, its own 4x-faster coding model

Cursor released version 2.0 of its AI code editor alongside Composer, a new in-house coding model claimed to be about 4x faster. The launch came up as evidence that developer products are being rebuilt agent-first, with speed and orchestration as the new battleground.

4x faster coding claimed

July 2025

Cursor (Anysphere)
Major Features & Updates

Cursor Agents on Web, Mobile & Slack

Cursor rolls out coding agents on web, mobile, and Slack

Cursor launched its AI coding agents on web and mobile with Slack integration, extending code agents beyond the editor window into ambient, always-on workflow software. The launch landed the same week Cursor poached key creators of Claude Code, making it product-strategy news as much as HR news.

March 2025

Cursor
Major Features & Updates

Claude 3.7 MAX

Cursor ships Claude 3.7 MAX mode

Cursor shipped Claude 3.7 MAX, a mode giving the agent the full context window and higher tool-call limits with Claude 3.7 Sonnet. It is aimed at harder, longer coding tasks at premium usage-based pricing.