Cursor adds Rollouts and Security Review
Teams and Enterprise gain deployment monitoring and PR security reviews. Rollouts can propose a revert or delegate a fix, but does not merge or roll back changes automatically.
/coding-agentscursor
Anysphere
Cursor
Coding environment with IDE and CLI agents, cloud execution, shared project context, code review, and enterprise controls.
IDECLICloudWebMobileCode reviewActiveProprietaryBYOKMCP Yes
Recent product updates
Reviewed
Selected releases and additions, with links to the original announcements.
Cursor Projects enters beta
A cloud coordinator plans work, delegates implementation, and shares context across agents. Projects can also respond to schedules and connected events. Availability began rolling out in beta.
Attributed output
GitHub public-attribution count, 2026-06-09
1.1MAttributed outputGitHub public-attribution count, 2026-06-09Commits and PRs on public GitHub provably produced by this agent. Exact if it is visible through its own bot or workflow, an estimate if measured through attribution it declares (these can swing 30 to 40 percent between measurement runs at multi-million scale), and a floor either way.Index rank · all agents
#3 of 40 CLI agents
#3Index rank · all agents#3 of 40 CLI agentsRanked by agent-opened PRs in the trailing 4 complete weeks, matching the index. Agents without a tracked PR series follow, ordered by total attributed output. Other filters or sorts change the list position.Business signal
Reported annualized revenue · Jun 2026
$4BBusiness signalReported annualized revenue · Jun 2026Reported by the press (Jun 2026), not officially disclosed or independently confirmed.Free Hobby / $20+ Individual
Entry price
Subscription and model usage
Free Hobby / $20+ IndividualEntry priceSubscription and model usageNot published
GitHub stars
No public repo
Not publishedGitHub starsNo public repoNot published
npm demand/wk
Public npm counter
Not publishednpm demand/wkPublic npm counterWeekly package downloads as an install-demand floor; adjusted when the vendor's main install channel is no longer npm.From the public record
What the public PRs reveal
452 PRs · 335 repositories
Search window: – · captured Oct 14, 2026 (UTC)
Cursor-attributed work gets CodeRabbit follow-ups
Separate study: CodeRabbit follow-ups, May 25–October 11, 2026
In CodeRabbit’s May 25–October 11 audit, we found 12 original PRs crediting Cursor, followed by 15 CodeRabbit auto-fix PRs; 11 follow-ups were merged. This cohort was selected through CodeRabbit, so it cannot tell us what fraction of this agent’s overall work receives follow-up fixes. Credits can name a model without identifying its client.
Specific work you can inspect
millionco/react-doctor: “feat(rules): add UI-library composition and correctness rules behind capability gates” (merged). rmcrackan/Libation: “Add HasSubtitle and TitleHasColon search fields” (merged).
This is a sample of publicly attributed work, not total usage or a market-share estimate. Repositories are not customers. Attribution is declared in metadata, not independently verified tool execution; it does not prove exclusive authorship or autonomy. PR status is recorded at capture; merged does not establish correctness.
Method and audit data
Up to 100 most recently created search results in each of five consecutive 28-day windows; post-filtered for attribution in the PR body, bot author, or Codex branch. Not a random or exhaustive sample.
Returned records dated Jun 21, 2026–Oct 11, 2026.
Download the dated audit evidence (JSON)Activity Over Time
Weekly attributed output on public GitHub. Output visible through the agent's own bot or workflow is exact; output measured through declared attribution is an estimate. Window: 2025-02-03 to 2026-10-12. The dashed span marks weeks when the vendor's attribution defaults were still changing, so commit levels there reflect how much work was signed, not how much work happened.
Market
Developer output
Public GitHub commits attributed to Cursor
Cursor's background agent declares attribution; its editor-driven edits land under the developer's identity and are invisible, so this is a floor. Cursor began attributing this work in early 2026 (almost nothing before February), so the series reflects that rollout rather than a sudden change in usage.
Last full week (ending 2026-10-18): 382K signed commits (+6.1% week over week), search estimate.
Distinct users: 3,944 humans opened the 12,600 most recently sampled PRs, about 3.2 PRs per person.
Merge rate: 71.1% of its pull requests get merged (574,247 of 807,888 over 8 settled weeks, 2026-08-10 to 2026-10-04). The window ends about ten days back so recently-opened PRs have had time to merge.
GitHub review activity: 6.9% of its merged PRs had at least one GitHub review record before merging; the other 93.1% had none. This is a broad weekly signal and can include bot reviews; the human-review read below comes from the enriched PR sample.
In significant repos (company-owned or 10+ stars), 77.8% of its merged PRs got a human review. Across all sampled repos, the rate is 35.8%. Human review means a non-author, non-bot GitHub reviewer in the PR-lifecycle sample.
Median time to merge: 1.02 h. The unreviewed median is 0.64h. The human-reviewed median is 1.84h. From a 120-PR sample merged 2026-07-30 to 2026-08-28.
Test-inclusion: 55.8% of its sampled merged PRs touch a test file. This is a rough read on production readiness. 2 large PRs hit GitHub's 100-file path read cap, so treat this as a floor.
Flow share: 1.8% of the 1999 most recent public PRs sampled (36 PRs, 2 windows, last 2026-09-27).
Task lens: 614K agent-opened PRs. Background-agent pull requests carrying hosted task metadata. Editor work opens PRs under the developer and is invisible.
Snapshot 2026-06-09, confidence Medium, tracked since 2026-06-09. Read the methodology.
- Index rank · all agents
- #3 of 47 · #3 among CLI agents
- Demand share
- No public npm package
- Business signal
- $4B Reported annualized revenue · Jun 2026
- GitHub share
- No public repo
- npm channel
- No public npm package
Best fit
AI-first IDE workflowsEnterprise IDE rolloutMulti-model teamsBackground agents
Direct competitors
- On June 16, 2026 SpaceX agreed to acquire Cursor (Anysphere) for $60B in an all-stock deal, days after SpaceX's IPO. It exercises the April 2026 option and is expected to close in Q3 2026 pending regulatory approval.
- Revenue roughly doubled in three months to about $4B (Forbes, June 2026), up from a reported ~$2B in February 2026; the strongest standalone IDE revenue signal in the index.
- About 700 employees; Cursor says it serves 60% of the Fortune 500 (June 2026).
- SpaceX was a training-compute partner before the acquisition: it supplies Colossus compute to scale Cursor's Composer models (announced April 2026) and plans to ship the jointly trained model in Cursor and xAI's Grok Build.
- Composer, Cursor's own coding models, are built on Moonshot open models; Cursor says Composer 2.5 (May 2026) is over 85% its own work, part of a push to cut reliance on frontier labs and control pricing.
- The strategy is a response to Anthropic dependence: at one point Cursor was a reported 40-50% of Anthropic's revenue, and Claude Code's launch (over $2.5B run-rate by February 2026) prompted a January 2026 internal push to build in-house models.
All sourced signals (4)ExpandHide
| Signal | Value | Type | Confidence |
|---|---|---|---|
| Reported ARR | Over $500M ARR reported in 2025 | Reported | Medium |
| Reported valuation | $9.9B reported valuation in 2025 | Reported | Medium |
| SpaceX model-training partnership | SpaceX and xAI supply Colossus compute to train Cursor's Composer models (Apr 2026) | Official | High |
| SpaceX acquisition option | SpaceX holds a call option to buy Cursor for $60B, or pay $10B to keep collaborating, through end of 2026 | Reported | Medium |
Where Cursor Works
Languages
TypeScript42%
Python21%
JavaScript13%
HTML7%
Rust4%
Go2%
Frameworks
Next.js21%
React17%
Flutter16%
Astro9%
React Native9%
Electron5%
Industries
AI/ML44%
Devtools7%
Gaming7%
Media5%
Crypto/Web35%
Finance5%
Countries
United States of America17%
China10%
United Kingdom9%
Canada6%
India6%
France5%
Share of Cursor's own classified sampled work by segment; countries reflect where its PR authors are based. Full breakdowns: Languages, Frameworks, Industries, Map.
Pricing And Economics
ExpandHide
Pricing snapshot: 2026-09-25. Check the vendor for current terms.
Hobby is free. Individual plans start at $20/month; Teams starts at $40/user/month; Enterprise is custom.
- Entry price
- Free Hobby / $20+ Individual
- Top self-serve price
- n/a
- Pricing unit
- Subscription and model usage
- Token transparency
- Token rate card
- Included usage
- Plans include model usage, with additional on-demand usage billed separately.
- Overage path
- Pro, Pro+, and Ultra have different allowances. Team features and model access depend on the tier.
Retail Token Equivalent
Allowance is not a fixed token entitlement. Consult the linked vendor pricing and your account’s usage limits.
| Plan | Price | Allowance | Unit | Token / Overage Notes |
|---|---|---|---|---|
| Hobby | Free | Limited agent use. | Subscription and model usage | |
| Individual | $20/mo | Starting price; Pro, Pro+, and Ultra have different limits. | Subscription and model usage | |
| Teams | $40/mo | Starting monthly price per user. | Subscription and model usage |
Unit Economics (Measured)
Mean API-retail cost per completed task on the Artificial Analysis coding agent index (358 tasks across 3 suites, pass@1 over 3 runs, snapshot 2026-06-09). These are pay-per-token API costs, not subscription effective costs.
| Harness | AA index | Cost / task | Tokens / task | Time / task |
|---|---|---|---|---|
| Cursor CLI, Composer 2.5 | 62.9 | $0.07 | 1.4M | 9m |
| Cursor CLI, Composer 2.5 Fast | 62.9 | $0.44 | 1.6M | 7m |
| Cursor CLI, Opus 4.7 (medium) | 61.2 | $1.47 | 1.5M | 8m |
| Cursor CLI, GPT-5.5 (medium) | 57.8 | $1.61 | 1.5M | 6m |
| Cursor CLI, GPT-5.4 (medium) | 52.2 | $1.53 | 2M | 8m |
| Cursor CLI, Composer 2 | 48.5 | $0.07 | 1.7M | 9m |
Break-even vs API retail on the top harness (Cursor CLI, Composer 2.5, $0.07/task): Individual at $20/mo covers about 286 tasks; Teams at $40/mo covers about 571 tasks. Assumes the Artificial Analysis task mix.
Summary checked 2026-09-25.
Hobby is free. Individual plans start at $20/month; Teams starts at $40/user/month; Enterprise is custom.
- Plans include model usage, with additional on-demand usage billed separately.
- Pro, Pro+, and Ultra have different allowances. Team features and model access depend on the tier.
Company And Funding
ExpandHide
- Ownership
- Private
- Founded
- 2022
- Headquarters
- San Francisco, CA
- Employees
- ~700
- Total raised
- ~$3.4B
- Most recent round
- Series D, $2.3B (Nov 2025)
- Round leads
- Accel, Coatue
Founded as Anysphere in 2022 by MIT classmates Michael Truell (CEO), Sualeh Asif, Arvid Lunnemark, and Aman Sanger. On June 16, 2026 SpaceX agreed to acquire Cursor for $60B in an all-stock deal, days after SpaceX's IPO and exercising an April 2026 option; the deal is expected to close in Q3 2026 pending regulatory approval. SpaceX had also been a model-training partner, supplying Colossus compute to scale Cursor's Composer models.
Notable investors
Thrive CapitalAndreessen HorowitzAccelCoatueDST GlobalNvidiaAlphabetOpenAI Startup Fund
Enterprise And Security
ExpandHide
- SSO / SCIM
- Yes
- Deployment
- Self-hosted
- Trains on your code
- Not trained on your code
- Zero data retention
- Yes
Compliance
SOC 2 Type II
Admin controls
SAML SSO / SCIMOrg policiesPrivacy Mode enforcementAdmin dashboard
Cursor added self-hosted machines and team worker pools on September 2, 2026, allowing tool execution inside customer infrastructure; this does not establish that model inference is self-hosted. Privacy Mode is on by default for Enterprise teams. Cursor holds zero-data-retention agreements with OpenAI, Anthropic, Google Vertex, and xAI; in Privacy Mode no code is stored or used for training.
Product
Providers
OpenAIAnthropicGoogleDeepSeekxAICursor
Model families
AutoGPTClaudeGeminiDeepSeekGrokCursor models
Capabilities
IDE agentCursor Agent CLIBackground agentsSide chatsCloud agent hooksCodebase searchMulti-model routingBYOKMCPBugbotRulesProjects (beta)Security ReviewRolloutsSelf-hosted machines
Tradeoffs
- Closed-source editor
- Cost depends heavily on chosen model and context size