@dijedont Tbh can't say yet. tinybar only sees tokens, not tasks.
But I built most of it with opus and code review, web-research, computer-use, etc with Sol
On $20 plans, $1 of Sol costs 7x what $1 of Opus does. A task, only 1.3x
Just the API equivalent usage for a subscription isn't enough. You need to show the efficiency of the models as well.
@AdamHoltererer Ah true! Compared by effort instead of score...
But still, lined up by score, codex leads by 3.2-4.7x
(The api $ each plan gets me comes from 1 week of my local logs though)
scores and cost per task are from @ArtificialAnlys.
the api $ each plan gives comes from my own token logs and limit readings, measured with tinybar
https://t.co/RqCBzQpW8h
Introducing Tinybar, a native Swift alternative to CodexBar.
• AI usage limits in your menu bar
• ~20 MB RAM & 0% CPU at idle
• Free & open source
• No telemetry. No bullshit.
https://t.co/RqCBzQqtXP
@solidoxx Already shipped an improvement.
More context in your gh issue: https://t.co/XG7rDJoYGn
Upgrade with `brew update && brew upgrade --cask thiloreintjes/tinybar/tinybar`
Lmk what you think
@solidoxx Turns out you just have a background color that is very close to the one of claude...
It's hard to see in your screenshot, but one can still see it.