Fast Premium Request is not fast

Discussion
1

After using Fast Premium Requests included in the Pro plan, I have to say—there’s nothing “fast” about it. The plan promises 500 Fast Premium Requests per month, but when using Sonnet 3.7 Thinking Model, the experience is painfully slow.

The biggest issue? Code generation speed. Even for minor edits, it takes an excruciating amount of time to get a response. The problem isn’t just latency—it’s the fact that Sonnet 3.7 Thinking Model struggles heavily with code generation itself, creating a major bottleneck.

Despite supposed performance improvements, workflow disruptions and productivity loss are unavoidable. As a paying customer, I can’t help but wonder: can this really be called “Fast” Premium Requests?

Is anyone else facing the same frustration?

8 Likes
2

I think there’s a bottleneck in upper management at Anysphere. Overloading the current AI model isn’t a problem the devs there can fix. My guess is the IT team is terribly micromanaged because they’ve implemented the latest tech to block bots, but the real issue is their Agent is burning through API credits too fast. They didn’t design the resources well, and their managers are blaming bots from China.