Cursor Pro is the standard paid tier for individual developers who use the AI code editor as their primary development environment. For $20 per month — or $16 per month billed annually — Pro removes the usage caps that limit the Hobby plan and unlocks the full power of Cursor's AI features. Unlimited Tab completions mean every keystroke benefits from AI predictions. The $20 monthly credit pool gives you access to frontier models like Claude Opus and GPT-5.4 for tasks that demand the most capable AI.
Pro also enables MCP server integration for connecting external tools, cloud agents for background task execution, and unlimited auto mode where Cursor selects the optimal model for each task at zero credit cost. Most Pro users find that auto mode handles the vast majority of their daily coding needs, reserving the credit pool for complex refactors, detailed code reviews, and multi-step agent sessions.
Start Free, Upgrade to Pro Compare All PlansA detailed walkthrough of what the Pro subscription unlocks beyond the free Hobby plan.
The Hobby plan limits the number of Tab completions you receive per day. Pro removes that cap entirely. Every keystroke triggers an AI prediction — multi-line function bodies, conditional logic, error handling patterns, API calls, and return statements. The completion engine uses frontier models with sub-200ms latency, running locally through the editor without any API round-trip for the suggestion display. Tab completions on Pro never consume credits from your monthly pool. They are fully unlimited and independent of the credit system. This single feature justifies the Pro price for most developers who write code in Cursor for more than an hour per day.
Pro subscribers receive $20 in credits at the start of each billing cycle. Credits fund two categories of usage: manual premium model selection and agent mode sessions. When you open the model selector and choose Claude Opus for a complex architectural question, or GPT-5.4 for a nuanced code review, the per-query cost deducts from your pool. Agent mode sessions — where Cursor autonomously writes code, runs commands, and iterates — also draw from credits when using premium models. Auto mode, which intelligently routes each task to the optimal model, is unlimited and free. Credits reset monthly. They do not roll over. The Cursor documentation publishes current per-model credit rates.
Model Context Protocol servers connect Cursor to external data sources. Query your production database schema, pull API documentation, fetch deployment status, or read from monitoring dashboards — all from within the chat panel. MCP servers give the AI real-time context beyond your local codebase. Cloud agents execute tasks on remote infrastructure while you continue coding locally. Kick off a test suite, deploy to staging, or run a migration, and the cloud agent reports results back to your Cursor session. Both features are exclusive to paid plans and fully included in Pro.
Pro unlocks manual selection of every model in the Cursor model roster. Claude Sonnet for fast, accurate code generation. Claude Opus for deep architectural reasoning and complex refactors. GPT-4o for broad general-purpose coding. GPT-5.4 for the latest frontier capabilities. Gemini for alternative perspectives and multimodal tasks. Each model has strengths that suit different task types, and Pro gives you the freedom to choose. Auto mode remains the recommended default — it selects the best model for each task without consuming credits — but manual selection is available whenever you need a specific model's capabilities.
See exactly where Pro sits between the free Hobby plan and the premium Pro+ tier.
| Feature | Hobby (Free) | Pro ($20/mo) | Pro+ ($60/mo) |
|---|---|---|---|
| Monthly Price | $0 | $20 ($16 annual) | $60 |
| Tab Completions | Limited daily | Unlimited | Unlimited |
| Monthly Credit Pool | None | $20 | $60 |
| Auto Mode | Limited | Unlimited | Unlimited |
| Agent Mode | Limited requests | From credit pool | 3x credit pool |
| Premium Models | None | All (from credits) | All (from credits) |
| MCP Servers | No | Yes | Yes |
| Cloud Agents | No | Yes | Yes |
| Composer | Limited | Full access | Full access |
| @Codebase Chat | Limited | Full access | Full access |
| VS Code Extensions | All | All | All |
| Student Discount | N/A | 1 year free (.edu) | No |
| Annual Billing | N/A | $16/mo | Available |
Pro serves most individual developers well. Here are the signals that indicate Pro+ would be a better fit.
If you consistently run out of your $20 credit pool before the end of the billing month, Pro+ triples that capacity to $60. Check your usage in the Cursor settings panel — it shows credit consumption by model and by day. Developers who use agent mode for 3 or more complex tasks daily typically benefit from the Pro+ credit pool. The switch is instant and prorated, so you can upgrade mid-cycle and receive the difference in credits immediately.
Agent mode consumes credits faster than manual model selection because it runs multiple queries per task — planning, executing, verifying, and iterating. A single agent session might use $1-3 in credits depending on task complexity and model choice. If you delegate feature builds, refactors, or debugging sessions to agent mode multiple times per day, the Pro credit pool can deplete quickly. Pro+ provides the headroom to use agent mode as a primary workflow tool rather than an occasional convenience.
Download Cursor free, explore with the Hobby plan, and upgrade to Pro when you are ready for unlimited completions and premium model access. The upgrade takes effect instantly with no interruption to your workflow. Students with a .edu email can activate one year of Pro at zero cost — every feature, no credit card required.
Download Cursor View All PlansAnswers about the Pro plan's features, credit system, pricing, and upgrade paths.
Unlimited Tab completions, a $20 monthly credit pool for premium models, unlimited auto mode, MCP server integration, cloud agents, full Composer access, and the complete VS Code extension ecosystem. Annual billing reduces the price to $16/month.
You receive $20 in credits each billing cycle. Auto mode is unlimited and free. Manually selecting premium models (Claude Opus, GPT-5.4) or running agent mode sessions draws from your pool at published per-query rates. Credits reset monthly and do not roll over.
Choose Pro if auto mode handles most of your work and you occasionally select premium models. Choose Pro+ ($60/mo) if you consistently exhaust your $20 pool — Pro+ provides 3x the credits for heavy agent mode users.
Claude Sonnet and Opus (Anthropic), GPT-4o and GPT-5.4 (OpenAI), and Gemini (Google). Auto mode selects the optimal model at zero credit cost. Manual selection draws from your $20 pool.
Yes. Sign up with a valid .edu email for one year of Pro at no cost. All features included, no credit card required. After the free year, continue at the standard rate or use the Hobby plan.