Can we please get some more information about how this rate-limited system is going to work?
Q1) Once we get rate-limited, is it just going to be the same as being in the current slow pool? Or will we not be able to make any requests?
Q2) How much does each model contribute to the maximum requests before getting rate-limited?
Q3) What is the rate-limiting limit actually going to be? (per hour, day, week etc)
Q4) can we try both new systems before deciding which one we want permanently? For example, one month on the new 500 requests system and then the next month on the new rate limiting system?
Overall, I’m worried that this is going to be a stealth nerf to the current 500 system.
Q1- yes you are trown into the slow-pool, where you can use MOST of the models that are allowed (newer models tend to be limitied to fast teir)
Q2- new models usually cost at least 1 tool call but I’ve seen some use .75 of a call.
Q3- limit is reset every month you started.
Q4- nah that’s already what it does use away until you use up all the 500 calls then it will just be slower and you’ll probably need to switch to using claude 3.7 for the rest of the time since claude 4 is in the fast pool only.
The rate limiting works on multiple time frames as far as I know (too many requests in an hour, too many in a day), but we’re able to offer limits high enough that very few people should ever see any rate limits to begin with.
Using more expensive models will get you close to that rate limit faster, so this is something to keep in mind!
2 & 3. Im hopeful we will be able to publish more information on this soon, but being a newly announced change that is coming to effect soon, we’re still working out the final details
I believe you should be able to choose which style plan your are on for a period of time. New users will likely be moved entirely to the new plan at some point, but we understand the desire to stay on the more predictable (but maybe more restrictive) 500 request allowance system.
regardless i still see that claude opus is sequestered to MAX only
will this removed for ultra? i can’t imagine keeping ultra and not having any access to OPUS unless i pay more
when i can have relatively unlimited with claude code in terminal
I choose to stay with my 497 fast requests. I have them, so I want to use them up. But my Pro plan got changed to the new one without my input. How do I revert it?
I’m now scared of using ANY request, even free models, so that I won’t hear later “you already used new Pro plan so you can’t revert back to old”
(For the record, I’m not a new user. My latest charge starts on June 6th.)
I think this is a downgrade, and a hidden one at that.
Because it would be silly to keep the option to return to the 500-request plan — that means the new change will reduce functionality and cause dissatisfaction, so they’re building in the ability to go back to 500 requests.
I don’t believe in generous offers for Pro users, especially when cheap models like DeepSeek are being limited to 60k context per request.
I think they’ve calculated everything and will offer around 17–20 requests per day to keep us within the same 500-request range. I think this is shrinkflation.
Or maybe some people can have a habit of using up all of their fast requests in a few hours? With the current system, it’s possible, but maybe with the new one, it will not be. Not sure though, but I guess today’s 500 requests do not have a rate limit baked in, so maybe some particular users might be affected by a new rate limit?
We’ll see I guess
