Review over T3


Beoordeeld met 1 van de 5 sterren

New Usage Limits destroyed great tool for good

I really loved the experience in this app. I signed up for 100 "expensive" queries (or 50 with search enabled) for $8, which imo is a really fair price. The experience in this app sometimes wasn't optimal (quality control or weird bugs that had be restart it a few times), but for $8, I really could take this.
However, since the latest usage limit updates, there's no such thing as fairness in the pricing any more. They introduced two scales, one "base" and the other "burst overage". Scale wouldn't be correct though, since you have no idea what a query takes from it or even what's the percentage you have left for either of those. For example, I just did a GPT 5.4 reasoning (rated as 3/3 expensive) search with 3 web searches enabled. The query took "3615" tokens and it ate away 3/4 of my base bar and 1/8 of my "burst overage" bar. The first one refills in 45 minutes, the second in about A MONTH!
This is just outrageous. Compared to the 100 "premium requests", this might be equivalent to maybe 4 points of the scale, which would have been 1/25 of my premium requests - which I'd be okay with. But now, all the transparency for the user is gone since I don't know how much a requests takes from my base & "burst overage" progress bar since there's nothing displayed in the user interface that indicates that. Oh an if you might think you can develop a gut-feeling of how much a prompt might take: you have to go to the settings (two clicks since the UI update last month or so) where the "scales" are visible. There's no other place to view them...
This just feels like the same as in every second mobile game with those in-game currencies, but *way* less transparent. And the best: they introduced a $50/month "premier" subscription 10xes the limits xD

I just feel tricked here. This is worse than anything I've ever experience on an online service. Additional to the usage limits, the UI got "updates" that makes it harder for users to select the reasoning effort and the amount of web searches, e.g. when you edit a already sent message in the chat. This is anti-UX and clearly aims to cut costs. The implementation of the search turn count incrementer also gives me the feeling of this.

Really a shame, I really liked this app for its features, but the $8 always felt too good to be true and this cash-grab seems to confirm this for me.

1 maart 2026
Review zonder uitnodiging