Introducing The X Subscription

I’m genuinely furious about this pricing update, and not because I misunderstood it, but because I understand exactly what it means for people who built their entire coding workflow around Cursor, only to watch a tool they relied on get wrapped into a larger package involving Grok, Grok Bot, and X Premium while the company presents it as something we should be excited about. Nobody running agents at 1am asked for social media perks, checkmarks, or fewer ads, because we paid for a coding tool and expected improvements to that coding tool, not a subscription that increasingly looks like a distribution channel for the broader xAI and X ecosystem. If that’s not what’s happening, then I’d genuinely love to hear why these unrelated services suddenly need to be bundled with something that was supposed to be a developer tool in the first place.

The “twice as much coding usage” claim is particularly insulting when you actually examine what it’s being compared against, because twice as much of what, using which models, and compared to which existing subscription? You keep comparing the new plans against Pro while conveniently ignoring Pro+, the $60 plan that many serious daily users actually depended on, and when people repeatedly ask what replaces it, the answer seems to be that there isn’t a direct replacement, Premium is aimed at newer users, Super is for heavy agent users, and existing subscribers should just stay on Legacy for now. That’s not a pricing improvement, that’s removing an important part of your offering and hoping nobody notices that the comparison looks considerably less impressive when you use the plan people were actually paying for. And then there’s the weekly shared usage pool, which you’re presenting as a benefit because supposedly one heavy day won’t consume an entire month’s allowance, except real development doesn’t happen in neat little weekly portions, and anyone who actually spends serious time debugging, refactoring, migrating systems, or running agents knows that some weeks are significantly heavier than others. A monthly allowance gave us the freedom to manage that variability ourselves, whereas a weekly limit simply introduces another artificial restriction, and people in this very discussion have already shown how quickly their percentages disappear during perfectly normal workloads. You trained people to rely on generous usage, combined Cursor, Grok, and Grok Bot into one shared meter, and now you’re marketing the resulting restrictions as though you’ve done everyone a favor.

And please stop pretending the third-party model situation is some minor detail, because staff have already confirmed that the new plans don’t include a third-party model allowance, meaning those models are charged at API list pricing from the first request while Legacy subscribers only retain their existing included allowances as long as they stay on their old plans. Saying those models are “still available” might be technically accurate, but it completely misses the point when the financial arrangement has fundamentally changed. Cursor built its reputation around model neutrality, giving developers one place where they could use whichever model worked best for a particular task without constantly worrying about additional charges, and that flexibility was arguably one of the biggest reasons people chose Cursor over competing tools. If Grok is supposed to become the preferred coding model, then let it earn that position by being better at coding, because pushing users toward it through pricing restrictions isn’t going to create loyalty, it’s going to make people seriously reconsider why they’re paying for Cursor at all. And the people reconsidering won’t necessarily be casual users, they’ll be the developers who kept the editor open all day, paid for higher tiers, and built entire workflows around the freedom that made Cursor attractive in the first place.

What makes this even more frustrating is the uncertainty surrounding Legacy subscriptions, because we’re being told that nothing changes today, there’s no announced migration timeline, and we’ll receive advance notice if anything changes, which sounds reassuring until you remember that the old subscriptions are no longer generally available and losing one could mean there’s no straightforward way back to the same pricing and usage conditions. That’s a pretty uncomfortable position for anyone who depends on this tool professionally, especially when the supposed replacement plans don’t offer the same combination of flexibility, model access, and predictable monthly usage. So no, the reaction in this thread isn’t confusion, and people aren’t upset because they failed to understand some brilliant new pricing structure. They’re upset because they understand perfectly well that a direct Pro+ replacement is missing, monthly flexibility is being replaced by weekly restrictions, third-party models are moving outside the included allowance, and coding subscriptions are being bundled with services developers never asked for, while the entire thing is being marketed as better value. If the strategy really is to integrate Cursor more deeply into the xAI and X ecosystem, then explain it openly instead of dressing up reduced flexibility as an improvement and expecting people to be grateful for perks that have absolutely nothing to do with their work.

I’m staying on Legacy for as long as it remains available, but don’t mistake that for unconditional loyalty, because I’m not married to your logo, I’m loyal to a tool that lets me work without having to watch a weekly percentage bar every time I have a demanding project. If Legacy eventually disappears without a genuinely comparable replacement, I’ll move to Codex, Anthropic, or whichever alternative gives me back the flexibility I’m paying for, and I suspect plenty of other heavy users will reach exactly the same conclusion. An X checkmark isn’t going to change that, a bundle of unrelated services isn’t going to change that, and calling weekly rationing a benefit certainly isn’t going to change that. Cursor became successful because developers trusted it to provide access to the best tools for their work, and watching that freedom gradually get replaced by ecosystem packaging and increasingly restrictive usage limits is an incredibly disappointing direction for a product that used to understand its audience so well. If I wanted rations, I’d move to the Soviet Union. At least they were honest about the shortages.