Cursor
AI code editor with a cheap in-house model; default model unstated and trains on code unless Privacy Mode is on.
Going away: OpenAI models in Cursor (Cursor may end access sooner), proposed for 12 Nov 2026, not confirmed. Source: help.openai.com, read 19 Oct 2026.
Should you use it?
Pick it if
People who want an AI code editor and are happy to choose the model themselves.
Best for: People who want an AI code editor and are happy to choose the model themselves (Coding agents Top 20)
Skip it if
You want a tool that names its default model; Claude Code does (Opus 5.5).
You give up: Trains on your code unless Privacy Mode is on; default model not named
The short version
- What this is
- A VS Code based AI code editor with chat, local and cloud agents, and PR review bots. Cursor became part of SpaceX on 14 Aug 2026.
- Why it matters
- Its own model is cheap per task, but the default model is unstated and personal code trains models unless you change a setting.
- Composer 2.5 in Cursor scored 62 on Artificial Analysis's index in May, at $0.07 a task s1.
- Turn on Privacy Mode if you don't want your code used for training s6.
- Since the SpaceX deal, Grok models sit in Cursor's own cheaper usage pool s3 s11.
Good at
- Its own Composer 2.5 model placed third of the agents Artificial Analysis tested in May 2026, at $0.07 a task s1.
- Lets you pick the top models by hand, including Claude Opus 5.5, at API rates from the included pool s3 s5.
- Ships often: remote control of local agents from iOS on 6 Oct, Security Review bot on 23 Sep 2026 s10.
Weak at
- Trains on your code unless you turn on Privacy Mode s6.
- Does not name its default model or what Auto routes to s3.
- OpenAI plans to stop supplying its models to Cursor; it has proposed 12 Nov 2026 as the end of the transition, and the date isn't final. After that, GPT models work only with your own OpenAI API key in local Chat and Agent (not Tab, Auto, cloud agents or the CLI) s12.
Our take
Cursor doesn't publish some things you'd want to know. Its docs don't say which model you get by default or where Auto sends requests s3, and its own Composer 2.5 isn't on arena.ai WebDev s4. If you pick Claude Opus 5.5 by hand you get the top WebDev model s4, billed at API rates from your included usage s5. The one independent test public, Artificial Analysis on 20 May 2026, found Composer 2.5 in Cursor third of its tested agents at 62, at $0.07 a task, close to the leaders at a small fraction of their cost s1. The live index values are not public s2. That is older than our 90-day window, so we haven't scored it.
The SpaceX deal closed on 14 Aug 2026 s11. The change users can see so far is in billing: Grok 4.5 to 4.7 now sit in Cursor's own model pool with Composer, exempt from the extra Cursor Token Rate s3. Grok 4.7 ranks 21 of 52 on arena.ai Agent s8. Cursor says the deal means more capable models at lower cost s11; no price change is listed yet. Personal code is used for training unless Privacy Mode is on s6, so turn it on if that matters to you.
Not scored yet
This is a First look: the call above comes from the maker's own pages, so there is no score yet. When an independent coding-agent benchmark publishes current results.
CodingAgents
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
Alternatives
Better or cheaper for coding, from our scores, then the ones we'd also look at. The same kind of product comes first.
- OpenCodeTop scoredCoding agent80FreeScores 80 in coding, first in our ranking; free. Side by side
- Claude CodeCoding agent70$20 a monthNames its default model, Opus 5.5, top of arena.ai WebDev; $20/month. Side by side
- GitHub Copilot (including Copilot CLI and cloud agent)Coding agentFirst look: Worth trying$10 a monthHalf the price at $10/month and works inside VS Code and JetBrains. Side by side
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 12
- s1Artificial Analysis: Cursor Composer 2.5 on the Coding Agent Index (20 May 2026)
- s2Artificial Analysis: Coding agents leaderboard
- s3Cursor docs: Models
- s4arena.ai Code Arena: WebDev leaderboard
- s5Cursor docs: Pricing
- s6Cursor: Data use
- s7Cursor pricing
- s8arena.ai Agent leaderboard
- s9Cursor: Security
- s10Cursor changelog
- s11Cursor blog: Cursor is now a part of SpaceX (14 Aug 2026)
- s12OpenAI Help Center: Using OpenAI models in Cursor
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 2 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- New plan prices from cursor.com/pricing (Plus $8, Premium $30, Super $100; third-party models billed separately). Added OpenAI's proposed 12 Nov 2026 end to supplying its models to Cursor.
- First look. Independent agent results are older than 90 days or test older models, so not rated yet.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.