I’ve been using Cursor IDE daily for over 5 months now, paying $20/month — which, living in Brazil, is not a small amount (over 120 BRL monthly). Overall, the tool itself is fantastic. Among all the AI-integrated IDEs I’ve tried, Cursor has by far the best and most seamless AI integration. It supports multiple cutting-edge models and is consistently updated with the latest advancements — something I deeply appreciate as a software developer working on complex codebases.
However, the way Cursor pushes its usage-based pricing model feels unfair and frustrating. For example, there’s a model called Claude 3.7 Sonnet Max, which is advertised as the smartest available, but it requires you to enable usage-based billing. This means additional charges on top of the $20 I’m already paying monthly. That alone feels like gated access to the real power of the platform.
To make matters worse, once I use up my 500 “fast requests”, I’m forced to wait several minutes for each prompt to be processed under the “slow request” mode. And during that wait time, Cursor prompts me with messages like: “Slow request, please wait or enable usage-based pricing.” — again pressuring me to spend even more.
What frustrates me most is that I’m already paying for access, and yet I don’t get full functionality unless I keep paying more. The only reason I pay for Cursor is to use the Claude 3.7 Sonnet model (not the “Max” version, which I avoid because of the added cost). If I only needed GPT or Gemini, I wouldn’t even pay — those models are free and unrestricted online.
In summary: Great product, frustrating pricing strategy. I wish Cursor would offer a more transparent and fair approach to its subscription tiers. For now, it feels like I’m paying a premium for a restricted experience, and that’s disappointing.