Skip to main content

t3 feels nerfed now

I really liked the older way that usage limits were computed as described here:

https://www.linkedin.com/posts/t3gg_yesterday-i-had-to-make-changes-i-didnt-activity-7293176522985676800-w3vH

1500 normal messaged & 100 for premium messages (not just Claude as mentioned above).

Today, I got this warning when using GPT 5.2 High Reasoning with just 2 messages with 5 web searches enabled.

You've used more than 75% of your usage limit. Your next reset is in 2h 3m. Consider upgrading.

This is, to say the least, disappointing.

Feature request: Please go back to the previous usage limit model (or at least something somewhat resembling it) - the one I agreed to when I initially signed up for t3 in March 2025.

Status: Closed3 comments

Log in to comment and vote

Comments3

  • Chocolate Pickle

    @Mark Florkowski That’s the whole point of t3 - being able to use the more expensive models for $8/month. I agree that there needs to be limits in place. There is a vast difference between a 2-message thread using the expensive models (causing 75% usage),vs 100 which is what the max-per-month used to be.

    When you say: For most users, the system allows much higher usage
    Is this anecdotal or do you have actual hard numbers to validate this statement?

    I still contend that I am not getting the same level of value that I once did in the past.

    Also, could you please consider making this thread public as I’d like to see if others feel the same way I do, if I am completely wrong on this or somewhere in between.

    Thank you.

    • Mark Florkowski

      Team

      | being able to use the more expensive models for $8/month

      this is still possible, you just can’t use those models in a way that costs way more than the cost of your subscription. If you are using short threads, and not attaching large files, you can get vastly more than 100 messages per month with these models.

      | Is this anecdotal or do you have actual hard numbers to validate this statement?

      We have analytics to confirm this.

  • Mark Florkowski

    Team

    For most users, the system allows much higher usage. Usage depletes faster if you use expensive models like Gemini Pro, Opus or GPT 5.2, have long threads, perform many web searches, or attach many files. That's due to the actual raw cost to us.

    These changes allow T3 Chat to be a more sustainable business, and continue to exist for the future, rather than running out of money because we were subsidizing usage