Honestly drives me crazy that CC doesnt show this on the bottom status bar a simple percentage “Weekly Quota Left: 99%” would be useful. Also noticed and idk if it was me or some mishap, a lot of my Claude sessions were not autocompacting, maybe its /loops fault not sure, but it made sense why I finally reached my weekly limit so insanely quickly recently.
I just want to know if my next prompt is about to eat shit at the last token. The number of times it like half works on something and falls apart halfway through.
You likely still wouldn't be able to figure that out even if you had a percent usage view available. Will your request take 1K tokens or 10K? Who knows! That's the magic of non-deterministic black boxes.
Doesn't it autocompact only if it is about to run out of context? On these newer models the context is 1M tokens so it's quite difficult to reach. I run compacts manually when I stop in a good place and tokens are over 300k
1m is too high, they used to autocompact a little too soon, maybe I tweaked defaults, it was kind of annoying, 200k tokens is too soon for me, I feel like the sweet spot is around 400k tokens.
Funnily enough Groks API charges less when you are below 400k tokens.
Half a million can get you reasonably far with enough of your codebase within the models context window.
I lowkey blame the 1m context window as the start of Anthropics worse woes earlier this year.
The Claude app has a usage indicator when you switch to code mode. I’ve been using this more and more now that their sandbox functionality has Improved.