I asked Hercules Agent or ran Debug Mode and it didn’t resolve my issue.
What is your issue about?
Hercules Agent: Something else
How is this affecting you?
A feature in my app isn’t working
What’s your issue?
More so a question but th option didn’t fit my question.
Has something changed with credit consumption? I feel like my credit usage has increased. I understand my app may have expanded which in turn credit usage with poorly optimised prompts but it feels larger than my app size. I have no idea about how it runs but it just feels significantly faster than the app growth.
What did you expect to happen?
So, I imported an app already built. The credit usage was a little high but expected dude to being a established app. I took a month off and coming back and it seems higher than before. Happy to be wrong but it just seems off. Probably just me and being poorly optimised requests.
Hey Liam,
Short answer: The old “Expert” model is now the “Balanced” model. Credits are priced and billed exactly the same as when you started, and there was no pricing change while you were away. I checked your account directly to confirm it.
The reason it feels higher is not app size, it is the model mix. In June a large share of your requests ran on a lighter, cheaper tier (Lite). Since you came back, your work has been running entirely on Expert, the most powerful and most expensive tier! Separately, the chats you have run lately are long, and a long chat costs more per message because the model re-reads the whole conversation on every step. Those two things together, not your app getting bigger, account for the jump. You can see the split yourself under Billing, then AI Usage.
Tips to bring it down immediately:
Lower the tier for routine work. In the model selector, switch from Expert to Balanced or Lite for everyday edits, and keep Expert for the genuinely hard problems. This is your biggest difference from before, since your recent work is all on Expert. (To clarify, our old Expert model is now the Balanced model. So if you’re used to THAT Expert, you should now be comfortable with Balanced!)
Start a fresh chat for each new task. The longer a single thread runs, the more each message costs, because the full history is re-read on every step.
Also if you want the lighter tiers to hold up on accuracy so you don’t need to reply on Expert, this guide should help: Make Hercules more accurate - Hercules
Thanks for the post! You can mark my reply as the solution if you felt it was helpful!