Repository navigation
Discrepancy between different opencode go usage dashboard #38255
Description
Description
AT midnight my time on the 22, all my models stopped working claiming weekly usage exceeded. When i looked at monthly limit dashboard i could confirm its 100%, but then when i looked in the granular usage dashboard, it showed i used barely 10$ of credit during the same timeframe.
i can confirm i didnt worked this week end so its kind of odd, i feel the 50% usage cap for the mimo 2.5 PRO model was applied over my whole 60$ account ?
Plugins
na
OpenCode version
na
Steps to reproduce
go in my account
check each dashboard
Screenshot and/or share link
No response
Operating System
windows 10
Terminal
na
Reacted by Amjad Ben Hedhili
Metadata
Metadata
Assignees
Labels
No labels
Activity
wywdemon commentedon Jul 22, 2026
I encountered the same problem. I didn't work on the weekend, but my monthly usage increased by about 20%
wywdemon commentedon Jul 22, 2026
My usage intensity is very low
GrandpaEJ commentedon Jul 23, 2026
Same problem here
And yes, i dont use that much
wywdemon commentedon Jul 24, 2026
Opencode Go has reduced the credit limit for dpv4 pro, so it's better to use Flash
wywdemon commentedon Jul 24, 2026
Why not use the free one
rushi1105 commentedon Jul 27, 2026
Same bug. Monthly shows 100% blocked (20+ day wait) but granular total is only ~$34. Rolling and weekly are at 0%. Monthly cap is $60 I've used nowhere near it.
diszone commentedon Aug 27, 2026
Same bug. I suspect it is an intentional design.
GrandpaEJ commentedon Aug 27, 2026
@diszone @MrMushrooooom @wywdemon @PiouPiou82
Guy's Look at the DOCS
they changed pricing .
like if u use deepseek/kimi model then you got only 15$ monthly
and if u use mimo-v2.5 then you got 60$ monthly
Davidli2013 commentedon Aug 28, 2026
Same problem on my OpenCode Go account.
DeepSeek V4 Flash hit its $30/month sub-cap → the whole plan is now locked. Console shows "每月用量 100.2%" but the granular breakdown is only ~$34 total — nowhere near the documented $60 monthly limit. The other models (V4 Flash Vision Exp $15, GPT 5.6 Luna $15, V4 Pro $15) are completely unused but can't be called either.
This matches the reported pattern: a single model's sub-cap saturates the whole account and blocks all models, even though per-model allowances should be independent. Also, cache-read tokens are billed server-side but invisible in local metering — see #41976.
Can a maintainer confirm the intended behavior, or reset the erroneous monthly meter? Happy to provide
opencode stats --days 30 --modelsandopencode export --sanitizeevidence.GrandpaEJ commentedon Aug 28, 2026
@Davidli2013 hey read this page
https://opencode.ai/docs/go/
PiouPiou82 commentedon Sep 1, 2026
they unban me in 1 days, had to take off the mention, might have been a weird transient bug.
I just stopped using any other model than those with 60$ credit mention, its just buggy as hell, it doesn't leave you the rest of the account ? wtv for real
they are super efficient in token consumption, i get like 5x less token burned when i route on opencode go than any other provider. with the 6x credit boost ... ill take the bug lol
tried using a model cheaper than mimo on the nous platform, glm 5.3 flash its .06 in/ .20 out right now. i burned 10$ in a day ... i never bust my 15$ subscription on opencode go with the EXACT same workflow .. it make no sense.
staying with them haha
AmjadHD commentedon Sep 30, 2026
Last 30 days usage (first time subscription):
DeepSeek v4.1: flash 2.03$
GPT-6-luna: 8.03$
I apparently reached luna's weekly limit, but now all paid models are blocked.
Tarquinen commentedon Oct 20, 2026
We're tracking this in #42985, so I'm closing this as a duplicate. Please follow that issue for updates.
— from 𝕺𝖕𝖊𝖓𝕮𝖔𝖉𝖊