Mikey

Investor and developer

[Mikey @biorevolution_]

@sama @sama Love the model, but the published 75–450 local messages per 5 hours doesn’t match what many Pro users are actually seeing. Some are hitting the limit after roughly 10 prompts on High. That gap needs an explanation.

[Mikey @biorevolution_]

@thsottiaux I’m not sure that holds up in practice. 5.5 XHigh still performs better than Sol Medium and cheaper, while Sol High hit the 5-hour limit in ~10 prompts. Repeated same-prompt tests also show Codex on WSL using far more quota than Windows. I have the full token breakdown if useful.

[Mikey @biorevolution_]

@thsottiaux I’m not sure that holds up in practice. 5.5 XHigh still performs better than Sol Medium and cheaper, while Sol High hit the 5-hour limit in ~10 prompts. Repeated same-prompt tests also show Codex on WSL using far more quota than Windows. I have the full token breakdown if useful.

[Mikey @biorevolution_]

@antirez Nice results. Have you tried GPT-5.6 Sol extensively in Codex? I'm curious about your usage. Many Pro users are seeing the 5-hour and weekly limits drain much faster than GPT-5.5 on similar workloads. The benchmark numbers suggest the opposite. Have you noticed anything similar?

[Mikey @biorevolution_]

$200 Pro is hard to justify when GPT-5.6 Sol on High can burn through the 5-hour limit in ~2 hours and gut the weekly quota. Benchmark data shows lower cost, fewer output tokens and fewer steps than 5.5, yet usage drains far faster. Banked resets don’t fix a work-blocking limit.

[Mikey @biorevolution_]

@reach_vb How am I supposed to use it when Pro hits the 5-hour limit in about 2 hours and my weekly quota drops from 100% to 85%? I've never used Ultra, only High, yet usage drains far faster than GPT-5.5 XHigh. Many Pro users are seeing the same thing, with no official clarification yet

[Mikey @biorevolution_]

@abraibrai @jaybinpark Nice job! Just wondering if the team is already aware that many Pro users are reporting GPT-5.6 is draining both the 5-hour and weekly limits much much faster than GPT-5.5. It seems to be affecting a lot of people.

[Mikey @biorevolution_]

@ChrissGPT I'm on the $200 Pro plan, and it's the same here. GPT-5.6 is chewing through the weekly quota about 10–15× faster than GPT-5.5 with the same usage. That can't be intentional.

[Mikey @biorevolution_]

@ajambrosino A lot of people are complaining about the absurd usage limits. Something clearly isn't right. We were using around 10× less of our weekly quota with GPT-5.5 under the same workload. At this point, it's hard to justify paying $200/month for Pro.

[Mikey @biorevolution_]

@reach_vb A lot of people are complaining about the absurd usage limits. Something clearly isn't right. We were using around 10× less of our weekly quota with GPT-5.5 under the same workload. At this point, it's hard to justify paying $200/month for Pro.

[Mikey @biorevolution_]

@sama Sam, a lot of people are complaining about the absurd usage limits. Something clearly isn't right. We were using around 10× less of our weekly quota with GPT-5.5 under the same workload. At this point, it's hard to justify paying $200/month for Pro.

[Mikey @biorevolution_]

@ajambrosino Hey Andrew, are you guys already aware of an issue with the usage limits? GPT-5.6 seems to be draining the weekly quota dramatically faster than GPT-5.5. With the same kind of usage, I used to consume roughly 10–15× less on 5.5, so this doesn't seem expected.

[Mikey @biorevolution_]

@ZerotoVibecode @thsottiaux I don't think that's how it works. Usage limits aren't supposed to fluctuate based on how many other people are using the service. If they did, everyone's quota would be completely unpredictable. This looks more like a bug than any kind of normal limit adjustment.

[Mikey @biorevolution_]

@thsottiaux Tibo, something seems off with the usage quotas. How is it possible that in just 40 minutes on the Pro plan I went from 50% of my weekly limit to 39%? With 5.5, it would have dropped by only 0–1%..

[Mikey @biorevolution_]

@thsottiaux @_simonsmith Tibo, something seems off with the usage quotas. How is it possible that in just 40 minutes on the Pro plan I went from 50% of my weekly limit to 39%? With 5.5, it would have dropped by only 0–1%.

[Mikey @biorevolution_]

@kevinwhinnery Can't wait! One question about the current 50% limit on Fable, do you expect it to be temporary? At launch, you said the goal was to eventually keep Fable permanently included in Max and the other plans. Is that still the plan?

[Mikey @biorevolution_]

@AnthropicAI Where are the remaining 11 days with the 100% weekly limits we were promised? On top of that, you made us wait almost three weeks, only to give us 6 days at just 50% capacity?

[Mikey @biorevolution_]

@trq212 Where are the remaining 11 days with the 100% weekly limits we were promised? On top of that, you made us wait almost three weeks, only to give us 6 days at just 50% capacity?