Evaluation record · cursor-agent

Cursor Agent

v3.x

Anysphere (SpaceX acquisition announced 2026-06-16, expected to close Q3 2026)

Agentideagentic-codingparallel-agentsproprietary

Adequate

About This Agent

Agent mode of Cursor, Anysphere's AI-native IDE. The product's primary surface is now agentic: parallel local agents, cloud/background agents running in isolated VMs, and the in-house Composer model line (Composer 2.5) alongside Claude, GPT, and Gemini. Anysphere IPO'd on Nasdaq in June 2026 and days later agreed to a $60B all-stock acquisition by SpaceX (announced 2026-06-16, expected to close Q3 2026 under its xAI subsidiary).

Last Evaluated: July 9, 2026

Official Website

Trust Vector Analysis

Dimension Breakdown

🚀Performance & Reliability

task completion accuracy

Assessment of agentic edit and task completion quality across Composer and frontier model options, drawing on vendor benchmarks and user reports

Evidence

Cursor Blog - Composer 2 — Composer 2 (2026-03-19) improved agentic coding accuracy and speed over Composer 1, which shipped with Cursor 2.0 in October 2025

mediumVerified: 2026-07-09

tool use reliability

Review of agent tool loop reliability across edit, search, terminal, and MCP integrations

Evidence

Cursor Documentation - Agent — Agent reliably uses codebase search, terminal commands, file edits, lints, and MCP tools within the IDE loop

highVerified: 2026-07-09

multi step planning

Evaluation of plan construction and multi-file refactoring on complex tasks in the agent-first interface

Evidence

InfoQ - Cursor 3 Agent-First Interface — Cursor 3 (April 2026) redesigned the IDE around agent plans and task orchestration as the primary workflow

mediumVerified: 2026-07-09

memory persistence

Review of rules, memories, and codebase indexing persistence

Evidence

Cursor Documentation - Rules and Memories — Project rules, user rules, and Memories persist conventions and learned context across sessions

mediumVerified: 2026-07-09

error recovery

Assessment of automatic iteration on lints, build errors, and failing tests

Evidence

Cursor Documentation - Agent — Agent reads linter and test failures and iterates automatically; loop detection prevents most runaway retries

mediumVerified: 2026-07-09

agent collaboration

Review of parallel and background agent orchestration capabilities

Evidence

InfoQ - Cursor 3 Agent-First Interface — Parallel agents run concurrently on separate tasks, locally via worktrees or in cloud VMs as background agents

highVerified: 2026-07-09

🛡️Security

tool sandboxing

Security review of cloud VM isolation versus local execution approval model

Evidence

Cursor Documentation - Background Agents — Cloud/background agents execute in isolated VMs; local agents run terminal commands on the user's machine with configurable approval and sandbox settings

mediumVerified: 2026-07-09

access control

Review of org-level controls, SSO, and repository scoping

Evidence

Cursor Security — Teams/Enterprise offer SSO, admin controls, and centrally enforced privacy mode; repository access scoped via GitHub app for cloud agents

mediumVerified: 2026-07-09

prompt injection defense

Threat surface analysis of untrusted content ingestion against documented mitigations

Evidence

Cursor Security — Agents process untrusted repo content, web results, and MCP outputs; command allowlists and approvals mitigate but injection hardening is not publicly detailed

lowVerified: 2026-07-09

data isolation

Review of privacy mode guarantees and tenant isolation for cloud agents

Evidence

Cursor Security — Privacy mode guarantees code is not stored or trained on; SOC 2 attested infrastructure with per-tenant separation for cloud agents

mediumVerified: 2026-07-09

open source transparency

Source availability assessment

Evidence

Cursor — Proprietary VS Code fork; Composer models and agent harness are closed source

highVerified: 2026-07-09

🔒Privacy & Compliance

data retention

Review of privacy mode retention guarantees and default settings

Evidence

Cursor Privacy — Privacy mode (default for Teams) prevents code storage and training use; without it, telemetry and code data may be retained

mediumVerified: 2026-07-09

gdpr compliance

Compliance documentation assessment

Evidence

Cursor Security — SOC 2 Type II attestation, DPA availability, and GDPR-aligned processing terms for business customers

mediumVerified: 2026-07-09

third party data sharing

Data flow analysis across multi-provider model routing

Evidence

Cursor Documentation - Models — Prompts and code context are routed to the selected model provider (Anysphere Composer, Anthropic, OpenAI, Google), expanding the data-processing surface

mediumVerified: 2026-07-09

local deployment option

Deployment options assessment

Evidence

Cursor Documentation — IDE runs locally but all model inference is cloud-based; no on-premises or fully offline mode

highVerified: 2026-07-09

👁️Trust & Transparency

documentation quality

Documentation completeness review

Evidence

Cursor Documentation — Thorough docs covering agent mode, background agents, rules, MCP, models, and enterprise administration

highVerified: 2026-07-09

execution traceability

Review of action visibility, diff review workflow, and agent logs

Evidence

Cursor Documentation - Agent — Every agent action is visible: diffs reviewable before apply, terminal output streamed, and per-agent activity logs in the agent-first UI

highVerified: 2026-07-09

decision explainability

Assessment of plan narration and change rationale quality

Evidence

InfoQ - Cursor 3 Agent-First Interface — Agents narrate plans and rationale in the task view, though depth of explanation varies by selected model

mediumVerified: 2026-07-09

open source code

Open source assessment

Evidence

Cursor — Closed-source editor fork and proprietary Composer models; no published weights or harness internals

highVerified: 2026-07-09

community activity

Community engagement and release cadence analysis

Evidence

Cursor Community Forum — Very large active user community, busy forum, and rapid release cadence (Cursor 2.0 Oct 2025, Composer 2 Mar 2026, Cursor 3 Apr 2026, Composer 2.5 mid-2026)

highVerified: 2026-07-09

⚙️Operational Excellence

ease of integration

Onboarding and integration friction assessment

Evidence

Cursor Documentation — VS Code fork imports existing extensions, settings, and keybindings; developers are productive in minutes

highVerified: 2026-07-09

scalability

Scalability assessment of parallel and cloud agent execution

Evidence

Cursor Documentation - Background Agents — Cloud agents offload long tasks to isolated VMs and run in parallel, decoupling agent throughput from local hardware

mediumVerified: 2026-07-09

cost predictability

Pricing model analysis including the June 2026 Teams restructure; flat tiers are clear but usage-based overages reduce predictability for heavy agent users

Evidence

Cursor Pricing — Free / Pro $20 / Pro+ $60 / Ultra $200 per month; Teams $40 per seat. Tiers include usage allowances with overage billing for heavy frontier-model use

Cursor Blog - Improvements to Teams Pricing — June 2026 Teams overhaul (effective 2026-07-01 for renewals): Standard seat $40/mo ($32 annual) with split usage pools for Composer/Auto vs third-party API models; new Premium seat at $120/mo ($96 annual) with 5x usage; Cursor estimates lower costs for ~90% of teams

highVerified: 2026-07-09

monitoring capabilities

Monitoring and admin analytics features assessment

Evidence

Cursor Documentation - Teams — Team usage dashboards and admin analytics; deep observability of agent behavior requires external tooling

mediumVerified: 2026-07-09

production readiness

Product maturity and adoption assessment; score held at 85 as strong adoption and revenue are offset by pending ownership change

Evidence

InfoQ - Cursor 3 Agent-First Interface — Mature, widely adopted product with massive enterprise and individual user base and sustained rapid iteration through Cursor 3

TechCrunch - SpaceX to acquire Cursor for $60B — Annualized revenue reached ~$4B by early June 2026; Nasdaq IPO June 2026 followed by $60B all-stock SpaceX acquisition agreement (close expected Q3 2026), introducing ownership-transition uncertainty alongside strong commercial traction

highVerified: 2026-07-09

Strengths

+Agent-first IDE redesign (Cursor 3, April 2026) makes agent orchestration the primary workflow
+In-house Composer model line (latest Composer 2.5) delivers very fast agentic edits alongside Claude/GPT/Gemini choice
+Parallel agents and cloud background agents in isolated VMs scale work beyond one task at a time
+Human-in-the-loop by design: reviewable diffs, command approvals, and visible terminal output
+Seamless VS Code compatibility for extensions, themes, and keybindings
+Privacy mode with no-storage/no-training guarantee, enforceable org-wide

Limitations

!Closed-source editor and models limit independent security and behavior auditing
!No offline or on-premises inference; all model calls go to cloud providers
!Usage-based overages above tier allowances make heavy agent usage costs less predictable
!Local agent terminal execution depends on user-configured approvals; misconfiguration widens risk
!Multi-provider model routing complicates data governance reviews
!Rapid release cadence occasionally introduces regressions and workflow changes
!Pending SpaceX/xAI acquisition (announced 2026-06-16, closing Q3 2026) creates governance, roadmap, and data-stewardship uncertainty for enterprise buyers

Metadata

license: Proprietary

supported models

0: Composer 1/2/2.5 (Anysphere in-house)

1: Anthropic Claude

2: OpenAI GPT

3: Google Gemini

programming languages

0: All languages supported by VS Code ecosystem

deployment type: Local IDE (VS Code fork) + cloud agents in isolated VMs

tool support

0: Codebase search and indexing

1: Terminal execution

2: MCP servers

3: Browser/web context

4: GitHub integration for cloud agents

first release: 2023 (IDE); Cursor 2.0 with Composer Oct 2025; Cursor 3 April 2026

pricing: Free / Pro $20 / Pro+ $60 / Ultra $200 per month; Teams (June 2026): Standard $40/mo ($32 annual), Premium $120/mo ($96 annual) per seat

company: Anysphere (Nasdaq IPO June 2026; ~$4B ARR; $60B all-stock SpaceX acquisition announced 2026-06-16, expected close Q3 2026 under xAI subsidiary)

Use Case Ratings

code generation

Best-in-class agentic IDE experience; parallel agents, fast in-house Composer models, and frontier model choice

data analysis

Strong for building and iterating on analysis code and notebooks, though not an analytics product itself

research assistant

Useful for technical research within codebases and docs; general research is outside its design focus

education

Visible agent reasoning and diffs help learners understand changes; generous free tier lowers the barrier

Similar Agents

Claude Code

Anthropic

GitHub Copilot Coding Agent

GitHub (Microsoft)

OpenAI Codex

OpenAI

Devin

Cognition