Grok 4.7 has become noticeably slower over the past few weeks
Has anyone else noticed that Cursor has become significantly slower when using Grok 4.7 over the past few weeks?
What makes this particularly frustrating is that Grok 4.7 seems noticeably slower than other models, even though those models consume API quota while Grok is included.
What is going on? Is this a known issue, a capacity problem, or something else?
I’m the only person among my peers who hasn’t switched to Claude yet. I’ve stuck with Cursor because I genuinely prefer the product and want to keep using it. But if the included Grok model continues to be this slow compared to alternatives that cost extra quota, it’s becoming increasingly difficult to justify staying.
It would be a real shame, but if this doesn’t improve, I don’t see much reason not to switch to Claude like everyone else.
Cursor team, could you clarify what’s happening and whether improvements are planned?
All the responses from official Cursor team members I could find on this forum seem to address isolated incidents or specific days with unusually high demand.
What I want to understand is the bigger picture: what is Cursor’s official position on the widespread reports of unsatisfactory performance that many users seem to be experiencing?
I’m on the Ultra plan, which advertises “Priority access to models during peak times”, but honestly, I haven’t noticed any benefit from this whatsoever.
Is this performance issue being acknowledged and actively addressed, or should we expect this to be the new normal for included models like Grok 4.7?
I think paying customers deserve a clear answer beyond explanations about temporary incidents or individual cases.
Turn Fast back on in the model picker if you’ve switched it off. It’s the default on Ultra and streams noticeably faster, at twice the per-token cost.
Use medium effort for routine tasks, since high and xhigh think longer per step.
@ckpulver (and @studev, if a run still feels unusually slow), a Request ID from a slow chat helps us check: open the chat’s … menu and choose Copy Request ID (in the Agents window it’s under Copy).
Hey Mohit, thanks for the info on the model settings, this will help.
To clarify my question: Is the Grok model’s current slow speed being addressed in development, or is this how we should expect it to perform going forward?