# Cursor vs Devin (November 2026): prices, agents and who each fits

> Cursor's plans pay for its own Grok and Composer models and bill Claude or GPT on top; Devin's $20 Pro plan includes frontier models and cloud agents.

URL: https://claudegpt.si/news/cursor-vs-devin  
Published: 2026-11-16T06:59:50.953Z  
Author: ClaudeGPT Si Newsroom  
Section: Reviews  
Topics: Cursor, Devin, Cognition, Coding agents, Comparison, Pricing  
Publisher: ClaudeGPT Si

## Key takeaways

- As of November 16, 2026, Cursor's individual plans cost $0 to $200 a month and Devin's cost $0, $20 or $200 a month.
- Cursor's included usage covers only its own Grok and Composer models, while Devin's Pro and Max quotas include frontier models from OpenAI, Anthropic, Google and SpaceXAI.
- Each company publishes scores only on its own benchmarks, so there is no shared test that ranks the two agents directly.

Cursor vs Devin comes down to where the work happens. Cursor is an editor first, with cloud agents added, and new customers now buy it inside The X Subscription, which runs from $8 to $200 a month as of November 16, 2026 ([Cursor pricing](https://cursor.com/pricing)). Devin is Cognition's cloud agent first, with a desktop editor added (Devin Desktop, the product that was Windsurf until June 2, 2026), and its individual plans cost $20 or $200 a month as of November 16, 2026 ([Devin pricing](https://devin.ai/pricing)).

The bigger difference is in what the money covers. Cursor's included usage pays for Cursor's own models (Grok and Composer) only, and Claude, GPT and Gemini are billed on top at API list price. Devin's paid plans include frontier models from OpenAI, Anthropic, Google and SpaceXAI in the plan's quota.

This comparison is based on the products' docs, pricing pages and published benchmark results, read on November 16, 2026. It is not a hands-on test.

## Cursor vs Devin at a glance

| | Cursor | Devin |
| --- | --- | --- |
| Company | Anysphere, sold through X Subscription Corp | Cognition |
| Main surface | Desktop editor, plus cloud agents, CLI, web and iOS | Cloud agent (web, Slack, Teams, API), plus Devin Desktop and Devin CLI |
| Free plan | Yes, limited Agent and Chat | Yes, light quota, SWE-2 at slow speed, Tab and inline edits unlimited |
| Cheapest paid plan (as of Nov 16, 2026) | Plus, $8/month, Cursor models only, no cloud agents | Pro, $20/month, includes Devin Cloud |
| Cheapest plan with cloud agents (as of Nov 16, 2026) | Premium, $30/month | Pro, $20/month |
| Top individual plan (as of Nov 16, 2026) | Ultra, $200/month (50x Plus usage) | Max, $200/month (weekly quota, no daily cap) |
| Claude, GPT, Gemini | Not in the plan; API list price from prepaid credit, Premium and up | Included in the quota on Pro and Max |
| Own models | Composer 2.5, Grok 4.5 to 4.7 | SWE-2, SWE-2 Flash |
| Usage reset | Weekly pool, shared with Grok | Pro: daily and weekly; Max: weekly |
| Past the limit | Prepaid top-ups, optional Auto Top-up | On-demand credits at API pricing, which never expire |
| Parallel cloud work | "As many agents as you want" | Up to 10 sessions on Pro and Max, unlimited on Teams and Enterprise |
| Teams (as of Nov 16, 2026) | $30, $100 or $200 per seat/month | $80/month team fee plus $40 per full seat (see below) |
| Code review | Bugbot, Premium and up | Devin Review, usage-based; public PRs free |
| Own benchmark | CursorBench 4.0 | FrontierCode, plus public suites |

## How Cursor and Devin differ in daily use

Cursor starts from the editor. Its [docs](https://cursor.com/docs) describe it as "a coding agent for building ambitious software," and most work happens in the Agent panel on your machine. [Cloud agents](https://cursor.com/docs/cloud-agent) run the same agent in isolated virtual machines with your repos, dependencies and secrets. You can start them from the desktop app, the web, iOS, Slack, Linear, the API or a `@cursor` comment on GitHub, and they push a branch back for review.

Devin starts from the delegated task. The [Devin docs](https://docs.devin.ai) call it "an autonomous AI software engineer that can write, run and test code," and suggest a rule of thumb: if you could do the task in three hours, Devin can most likely do it. You hand it a Slack thread, a Linear or Jira ticket, or a prompt in the web app, and it works in its own machine with a shell, editor and browser.

The two have moved toward each other. Cognition [launched Devin Desktop on June 2, 2026](https://cognition.com/blog/introducing-devin-desktop) as "the next generation of Windsurf," with an Agent Command Center that manages local and cloud agents in one place. The [Devin Desktop changelog](https://docs.devin.ai/desktop/changelog) shows version 3.12.59 shipping on November 12, 2026.

Cursor assumes you are at the keyboard and sometimes send work away. Devin assumes the reverse.

## What Cursor costs (as of November 16, 2026)

New Cursor customers start on The X Subscription, one plan that covers Cursor, Grok and X. These prices come from [cursor.com/pricing](https://cursor.com/pricing) and [Cursor's models and pricing docs](https://cursor.com/docs/models-and-pricing.md), as of November 16, 2026:

| Plan | Monthly price | Annual price | Included usage | Third-party models |
| --- | --- | --- | --- | --- |
| Free | $0 | | Limited | No |
| Plus | $8 | $80/year | Base pool | No |
| Premium | $30 | $300/year | 4x Plus | API list price |
| Super | $100 | $1,000/year | 15x Plus | API list price |
| Ultra | $200 | $2,000/year | 50x Plus | API list price |
| Teams | $30, $100 or $200 per seat | | Matches the seat's tier | API list price plus $0.25 per million tokens |
| Enterprise | Custom | | Pooled | API list price plus $0.25 per million tokens |

Three things matter more than the sticker price:

- **Plus is a Grok plan.** It runs Agent with Cursor's own models only, at fixed settings, with no cloud agents, Bugbot, Auto or your own API keys. Cloud agents start at Premium, $30 a month as of November 16, 2026.
- **The pool covers Cursor models only.** Grok 4.7, 4.6, 4.5 and Composer 2.5 draw from a weekly pool shared with the Grok apps. Past the pool, Composer 2.5 costs $0.50 per million input tokens and $2.50 per million output tokens, and Grok 4.7 costs $2 and $6, as of November 16, 2026.
- **Claude, GPT and Gemini cost extra from the first request.** They are billed at API list price from prepaid credit. On Teams and Enterprise, eligible third-party requests add a Cursor Token Rate of $0.25 per million tokens (as of November 16, 2026).

Customers who signed up earlier can stay on [legacy Cursor plans](https://cursor.com/docs/account/pricing/legacy-cursor-plans.md), including Pro at $20 a month and Pro+ at $60 a month (as of November 16, 2026), which still include an allowance for third-party models. Switching to The X Subscription is permanent. Our [Cursor pricing guide](https://claudegpt.si/news/cursor-pricing-plans-weekly-pool) covers the plans in more detail.

## What Devin costs (as of November 16, 2026)

Devin replaced its old Core and Team plans with five self-serve tiers on [April 14, 2026](https://cognition.com/blog/new-self-serve-plans-for-devin), and self-serve usage is now priced in dollars rather than ACUs (Agent Compute Units). Enterprise contracts still use ACUs. These prices come from [devin.ai/pricing](https://devin.ai/pricing) and the [self-serve plans docs](https://docs.devin.ai/admin/billing/self-serve.md), as of November 16, 2026:

| Plan | Monthly price | Usage | Concurrent sessions |
| --- | --- | --- | --- |
| Free | $0 | Light quota, limited models, SWE-2 at slow speed | None listed |
| Pro | $20 | Daily and weekly quota, frontier models, free SWE-2, Devin Cloud | Up to 10 |
| Max | $200 | Much larger weekly quota, no daily cap | Up to 10 |
| Teams | $80 team fee plus $40 per full seat | Each full seat gets a Pro-sized quota; flex seats draw on shared credits | Unlimited |
| Enterprise | Custom | Custom, billed in ACUs | Unlimited |

The Teams price reads two ways. The pricing page lists "$80/month for team plan + $40/mo per full dev seat." The billing docs describe $80 as a monthly minimum that full seats count toward: two full seats cost $80, three cost $120, and with fewer than two seats the rest of the $80 is added as prepaid credit. Check the total at checkout.

Past the quota, extra usage is bought as on-demand credits at API pricing. The credits roll over and never expire. Two details can stretch the quota. A linked ChatGPT Go, Plus or Pro subscription can pay for GPT usage inside Devin. And on Teams, every member can run up to 2 cloud sessions at a time in Free mode, powered by SWE-2, at no cost.

## Benchmarks: no shared test yet

Cursor and Cognition each publish scores on their own tests, and neither runs the other's agent. That makes a direct score comparison impossible from their own pages.

Cursor's [CursorBench 4.0](https://cursor.com/evals), introduced on September 10, 2026, uses "ambiguous, multi-file tasks from real Cursor sessions." On it, Composer 2.5 scores 27.7% at an average of $0.68 per task, and Grok 4.7 at extra-high effort scores 46.3% at $6.01 per task. The top entry is Gemini 4 Carbon High at 60.4%. Cursor notes that small differences may not be statistically meaningful.

Cognition says [SWE-2](https://cognition.com/blog/swe-2), released on September 10, 2026, scores 50.0% on FrontierCode 1.1 Main, its own benchmark, against 50.9% for Fable 5.1 and 53.3% for GPT-6 Astra. On Terminal-Bench 2.1 it reports 92.8%. On the harder Terminal-Bench 4.0 it reports 27.3%, well behind Fable 5.1 at 55.8% and GPT-6 Astra at 57.9%. [SWE-2 Flash](https://cognition.com/blog/swe-2-flash), released on November 4, 2026, scores 24.0 on FrontierCode 2.0 and, by Cognition's numbers, runs 3x faster than SWE-2 at a fifth of the cost. It is free in Devin Desktop and Devin CLI through November 30.

All of these are the companies' own numbers. The public Terminal-Bench leaderboard did not show its rows when checked on November 16, so there is no independent score to set beside them. Both run the same third-party models, so the model you pick may matter as much as the product. Our [SWE-2 pricing article](https://claudegpt.si/news/cognition-swe-2-pricing) covers Cognition's model in more detail.

## Who each one fits

Cursor fits developers who spend the day in an editor and want an agent next to them. It is the cheaper start for anyone happy with Cursor's own models: $8 a month for Plus, as of November 16, 2026. For people who mainly want Claude or GPT it can cost more than Devin, because those models are billed on top of a $30 Premium plan or higher.

Devin fits people who want to hand off whole tickets and review pull requests later, and teams that live in Slack, Linear or Jira. At $20 a month for Pro, as of November 16, 2026, it is the cheaper way to get both cloud agents and frontier models in one quota, with a cap of 10 sessions at a time.

The fairest test is one real ticket from your backlog on each free tier, watching how much quota it uses and how much you have to fix. For other matchups, see our comparisons of [Windsurf vs Cursor](https://claudegpt.si/news/windsurf-vs-cursor-comparison) and [Codex vs Cursor](https://claudegpt.si/news/codex-vs-cursor).

## Frequently asked questions

### Is Devin better than Cursor?

It depends on how you work. Cursor is built around an editor with cloud agents added, while Devin is built around handing off whole tasks to a cloud agent, and no shared benchmark compares the two directly.

### Is Devin free?

Devin has a free plan with a light quota, limited models and SWE-2 at slow speed. Paid plans cost $20 a month for Pro and $200 a month for Max, as of November 16, 2026.

### Is Windsurf now Devin?

Yes. Cognition renamed Windsurf to Devin Desktop on June 2, 2026. Cognition says it is the same IDE, stays compatible with Windsurf settings and rules, and kept existing plans and prices unchanged.

### Can I use Claude in Cursor and Devin?

Yes, in both. Devin's Pro and Max quotas include Claude models, while Cursor bills Claude at API list price from prepaid credit on Premium ($30 a month as of November 16, 2026) and higher plans.

## Sources

- [Cursor pricing](https://cursor.com/pricing), Cursor
- [Models & Pricing](https://cursor.com/docs/models-and-pricing.md), Cursor
- [Legacy Cursor plans](https://cursor.com/docs/account/pricing/legacy-cursor-plans.md), Cursor
- [Cloud Agents](https://cursor.com/docs/cloud-agent), Cursor
- [Cursor docs](https://cursor.com/docs), Cursor
- [CursorBench evals](https://cursor.com/evals), Cursor
- [Devin plans and pricing](https://devin.ai/pricing), Cognition
- [Self-serve plans](https://docs.devin.ai/admin/billing/self-serve.md), Cognition
- [Introducing Devin](https://docs.devin.ai), Cognition
- [Devin Desktop changelog](https://docs.devin.ai/desktop/changelog), Cognition
- [Devin Desktop FAQ](https://docs.devin.ai/desktop/devin-desktop-faq.md), Cognition
- [Introducing Devin Desktop](https://cognition.com/blog/introducing-devin-desktop), Cognition
- [New self-serve plans for Devin](https://cognition.com/blog/new-self-serve-plans-for-devin), Cognition
- [Introducing SWE-2: Pushing the Pareto Frontier](https://cognition.com/blog/swe-2), Cognition
- [Introducing SWE-2 Flash](https://cognition.com/blog/swe-2-flash), Cognition
