Cursor CEO Issues Apology for Sudden Pricing Changes: What Pro Users Need to Know About the New Compute-Based Model
If you woke up to find your trusted AI coding assistant, Cursor, suddenly costing you more—or even cutting you off after just a few prompts—you’re not alone. In a move that caught much of the developer community off guard, Cursor’s CEO Michael Truell has publicly apologized for how the company rolled out major changes to its Pro plan pricing. The story is bigger than a single subscription fee: it’s about trust, transparency, and how fast-evolving AI platforms are redefining what “value” really means for their users.
Whether you’re a loyal Cursor user, a developer keeping tabs on AI productivity tools, or someone just wondering what all the fuss is about, this deep dive will explain everything you need to know about the Cursor pricing controversy—what happened, why it matters, and what it says about the future of AI-powered coding platforms.
The Curious Case of Cursor’s Pro Plan: From “Unlimited” to Uncertainty
Let’s set the scene. Cursor, a popular AI-powered coding companion, has built its reputation on giving developers fast, integrated access to cutting-edge models like Anthropic’s Claude, OpenAI’s GPT-4, and Google’s Gemini. For many, the $20/month Pro plan was a no-brainer: pay once, code more, and never worry about running out of requests.
But in mid-June 2025, everything changed.
What Sparked the Outrage?
Overnight, Cursor quietly shifted from a clear-cut request-based system (think “500 requests per month”) to a new, compute-based model. Instead of simply counting requests, Pro subscribers were given a $20 monthly credit at API rates—which varied wildly depending on which AI model you used and how complex your prompts were.
Suddenly, users who were used to getting hundreds of code completions per month found themselves burning through their entire allocation after just a few advanced queries. Worse, many were blindsided: the Pro plan’s description changed from “Unlimited” to “Extended,” and some users said they only realized the impact after hitting abrupt rate limits mid-project.
Here’s why that matters: For developers relying on AI to hit deadlines or ship features, running out of requests isn’t just an annoyance—it’s a blocker that can derail entire workflows.
Unpacking the Pricing Pivot: Why Did Cursor Make the Switch?
It’s easy to see why users felt frustrated. But from Cursor’s perspective, there’s a logic to the change.
The Growing Cost of Advanced AI Models
Modern AI models aren’t created equal. As tools like Claude Sonnet 4 and GPT-4.1 have grown more powerful, their computational demands—and the costs for companies like Cursor to serve them—have skyrocketed. In the old system, a simple code completion and a massive, context-heavy refactor both counted as “one request,” even though the latter could cost Cursor 100 times more in backend compute.
Cursor’s reasoning, according to CEO Michael Truell:
“We recognize that we didn’t handle this pricing rollout well and we’re sorry… But the new API-based pricing better reflects the actual computational costs of increasingly sophisticated AI models.”
— Michael Truell, CEO, Cursor
What Changed? Old vs. New Cursor Pro Plan
Here’s a breakdown to make sense of it all:
Previously (Request-Based): – $20/month Pro plan – 500 requests/month (regardless of complexity/model, with some exceptions) – Simpler, more predictable—great for budgeting
Now (Compute-Based): – $20/month Pro plan – $20 credit at API rates (number of queries varies with model and prompt complexity) – Roughly 225 requests with Claude Sonnet 4 – About 550 requests with Google Gemini – Up to 650 requests with GPT-4.1 – Unlimited usage only in “Auto” mode (Cursor decides which model runs your query) – Option to buy additional usage at cost, or set a spend limit
Key Point: The more powerful (and expensive) the model, or the more complex your task, the fewer prompts you can run before hitting your cap.
Why Did Users Feel Blindsided?
Let’s be honest: The tech world moves fast, but nothing riles up developers more than a sudden, poorly explained pricing change—especially when it impacts their daily workflow.
Communication Breakdown: The Core Issue
- Lack of Advance Notice: The shift was announced and implemented with little warning.
- Opaque Plan Description: The Pro plan went from “Unlimited” to “Extended” quietly, sowing confusion.
- No Clear Calculator: Users had no easy way to estimate how many prompts they could run with their typical usage.
Result?
Power users, especially those who relied on Anthropic’s Claude for complex coding tasks, found themselves running out of credits after just a handful of prompts. Imagine expecting a full tank for your road trip, only to realize you’ll run dry a few miles in.
Cursor’s Response: Apology, Refunds, and a Promise to Do Better
To its credit, Cursor’s leadership recognized the missteps—publicly.
The Official Apology
Michael Truell’s message to users was clear: “We recognize that we didn’t handle this pricing rollout well and we’re sorry.” (You can read more about the apology on TechCrunch and Cursor’s official forum.)
But words alone aren’t enough when trust is on the line.
Full Refunds Offered for Affected Users
Cursor is now offering full refunds for anyone who was unexpectedly charged or ran out of credits between June 16 and July 4, 2025. If you’re one of those users, you can email propricing@cursor.com to claim your refund.
Clarifying the Pro Plan
Cursor also took steps to clarify what’s actually included in the Pro subscription now: – Unlimited Usage of Tab and AI models in Auto mode (where Cursor routes your requests to available models) – $20/month “Frontier Model” Credit: This covers direct usage of premium models at API pricing – Option to Buy More at Cost: If you hit your cap and need more, you can pay for extra usage at Cursor’s cost
For power users, there’s now a new “Ultra” tier at $200/month, providing 20x more usage than Pro—ideal for teams or heavy-duty developers.
The Technical Nitty-Gritty: How Compute-Based Pricing Really Works
Let’s break it down with a real-world analogy.
Imagine you’re at an all-you-can-eat buffet. Under the old system, everyone got 500 plates—whether you filled them with salad or lobster. Under the new system, you get a $20 voucher. If you stick to salad (light queries, simple models), your voucher stretches far. But go for the lobster (complex code refactors with top-tier AI), and you burn through your credit fast.
How Quickly Can You Run Out of Credits?
- Simple queries (short prompts, small context) use minimal tokens, so you get more requests per dollar.
- Complex tasks (like refactoring an entire codebase, detailed AI code generation) can eat up 50,000+ tokens in one shot—meaning your monthly allocation could disappear with just a few operations.
- Model choice matters: Anthropic’s Claude is especially prized for coding, but it’s also pricier per token than GPT-4.1 or Gemini.
Here’s a quick cheat sheet for typical usage under the new Pro plan (these are rough estimates):
| Model | Approx. Prompts per $20/month* | |——————-|——————————-| | Claude Sonnet 4 | 225 | | Google Gemini | 550 | | GPT-4.1 | 650 |
*Actual number varies widely based on prompt complexity.
What’s Still Unlimited? The Role of Cursor’s “Auto Mode”
Feeling boxed in by credits? There’s a loophole—and it’s called Auto mode.
How Auto Mode Works
- When you use Auto mode, Cursor automatically selects which AI model to send your query to, based on factors like cost, capacity, and availability.
- Pro users still get unlimited usage in Auto mode, but you lose the ability to manually pick premium models like Claude Sonnet 4 for every task.
Here’s why that matters:
If you’re flexible about which AI model handles your prompts, you won’t hit hard limits. But if your workflow depends on a specific model’s strengths, you’ll need to keep an eye on your credit balance.
The Developer Community Reacts: Frustration, Confusion, and Cautious Optimism
The backlash was swift and passionate—especially on Cursor’s public forums and X (formerly Twitter).
Real User Stories
- “I burned through my entire month’s credit in a single afternoon refactoring my codebase with Claude. I had no idea it would be so expensive.”
- “Unlimited became ‘extended’ overnight. Feels like a bait-and-switch.”
- “Appreciate the apology and refund offer, but I wish there had been better communication in the first place.”
But not every perspective was negative. Some users acknowledged the underlying problem—AI models are expensive to run, and “true unlimited” is rarely sustainable.
What Cursor’s Pricing Shift Signals for the AI Tool Market
Cursor isn’t the first, nor will it be the last, AI-powered company to grapple with the realities of pricing advanced compute.
The Industry Trend: From Flat Rates to Usage-Based Pricing
- AI models are evolving fast, and their resource demands are outpacing old-school subscription models.
- More companies are moving to metered, API-based pricing—charging users based on what they actually consume, rather than flat limits.
- This can lead to greater fairness (pay for what you use), but also more complexity and unpredictability for end users.
OpenAI, Anthropic, and Google Cloud AI all have nuanced pricing models based on usage, not just flat monthly fees.
How to Navigate the New Cursor Pricing: Tips and Best Practices
Here’s how you can make the most of Cursor’s new pricing—without unpleasant surprises.
1. Monitor Your Usage Regularly
Cursor now provides more transparent usage tracking in your account dashboard. Make it a habit to check your balance, especially after heavy work sessions.
2. Leverage Auto Mode for Unlimited Access
If you don’t need a specific model for every prompt, stick with Auto mode to avoid hitting credit caps.
3. Budget for “Frontier Model” Usage
If you rely heavily on advanced models, plan your workflow accordingly—or consider upgrading to the Ultra tier if your usage demands it.
4. Ask for Help
If you believe you were impacted by the recent changes, don’t hesitate to email propricing@cursor.com for a refund or clarification.
5. Stay Informed
Cursor has pledged to communicate future changes more clearly. Follow their official blog, subscribe to updates, and stay active in the community forums.
Frequently Asked Questions (FAQ)
Q: What exactly changed in Cursor’s Pro plan pricing?
A: Cursor moved from a request-based system (500 requests/month) to a $20/month compute credit at API rates. The number of prompts you can run now depends on which model you use and how complex your queries are.
Q: Why did users run out of credits so quickly?
A: Advanced AI models like Claude Sonnet 4 are more expensive to run, especially for complex prompts. Some users used up their entire monthly allocation after just a few heavy prompts.
Q: Is anything still unlimited on the Pro plan?
A: Yes, usage of Tab and models in Auto mode remains unlimited for Pro subscribers. Direct usage of premium models is capped by your $20 monthly credit.
Q: Who is eligible for a refund?
A: Any Pro subscriber who was unexpectedly charged or ran out of credits between June 16 and July 4, 2025, can request a full refund by emailing propricing@cursor.com.
Q: Can I control or limit additional charges?
A: Yes, Cursor now allows you to set a usage cap or buy additional compute at cost once you reach your monthly limit.
Q: Where can I learn more about API pricing for different AI models?
A: Check out the official pricing pages for OpenAI, Anthropic, and Google Cloud AI for the latest rates.
Final Thoughts: Transparency, Trust, and the Road Ahead for AI-Powered Coding
Cursor’s pricing saga is a case study in the growing pains facing AI platforms—and their users. As models become more sophisticated (and expensive to run), the lines between value, fairness, and simplicity get blurry. Cursor’s apology and refund offer are steps in the right direction, but the episode is a reminder of just how vital clear communication is in a world where technology evolves faster than most of us can keep up.
The takeaway?
If you depend on AI tools, stay vigilant—about your usage, your costs, and the fine print. Ask questions, give feedback, and don’t be afraid to speak up if something feels off. At the end of the day, the platforms that listen and adapt will be the ones that earn your trust—and your business.
Want more insights on AI tools, developer productivity, and the future of coding?
Subscribe to our newsletter and join a community of smart, curious tech professionals who never stop learning.
External References: – TechCrunch: Cursor’s Pricing Controversy – Cursor Official Forum – OpenAI Pricing – Anthropic Console Pricing – Google Cloud Vertex AI Pricing
Discover more at InnoVirtuoso.com
I would love some feedback on my writing so if you have any, please don’t hesitate to leave a comment around here or in any platforms that is convenient for you.
For more on tech and other topics, explore InnoVirtuoso.com anytime. Subscribe to my newsletter and join our growing community—we’ll create something magical together. I promise, it’ll never be boring!
Stay updated with the latest news—subscribe to our newsletter today!
Thank you all—wishing you an amazing day ahead!
Read more related Articles at InnoVirtuoso
- How to Completely Turn Off Google AI on Your Android Phone
- The Best AI Jokes of the Month: February Edition
- Introducing SpoofDPI: Bypassing Deep Packet Inspection
- Getting Started with shadps4: Your Guide to the PlayStation 4 Emulator
- Sophos Pricing in 2025: A Guide to Intercept X Endpoint Protection
- The Essential Requirements for Augmented Reality: A Comprehensive Guide
- Harvard: A Legacy of Achievements and a Path Towards the Future
- Unlocking the Secrets of Prompt Engineering: 5 Must-Read Books That Will Revolutionize You
