I think Grok Bot usage is draining my credit pool. I’m seeing many requests hitting claude-opus-5-thinking-low and cursor-grok-4.6-high-fast. I’m using neither model on Cursor. When I stop the bot, my credit pool stops draining.
Same here, let’s hope they will notice. Grok Bot used 52M of Claude Opus tokens for me while using 38% Weekly usage. And used Grok High - Fast as well.
Did your Grok Bot spawn any cloud agents in Cursor to achieve some task? Because then those cloud agents that grok bot created would be using Cursor credits no? What do you see in Spending?
No cloud agents were spawned. I kept all activity within the scope of my laptop and the computer on Grok Bot. My first post has the spending summary while a Bot was active in the app. Usage is bouncing between that claude and grok model. 99% of my chats and cloud agents on Cursor are behind Grok 4.6 Extra High Fast.
Grok Bot has its own weekly usage pool on Pro+, but once that’s used up it keeps going and spills onto your regular Cursor balance: credits (referral/promo) first, then on-demand if enabled. That spillover is the drain you’re seeing, which is why stopping the bot stops it. The claude-opus-5-thinking-low / cursor-grok-4.6-high-fast entries are models Grok Bot picks server-side, so they show up even though you never selected them. This is expected behavior, not a metering error, and it’s documented here.
The reliable fix is to pause long-running bots as the weekly meter nears 100%. Setting a $0 on-demand limit (Dashboard → Usage/Billing) stops paid on-demand charges, but note it won’t stop credits from being consumed.
If you’d like us to review the specific usage on your account, email [email protected].
@mohitjain , so it’s definitely a bug if these models are being used while Grok Bot itself is only at 38% of the weekly usage limit? Ticket T-F21229 - please take a look.
