# Do I have a workflow issue? 20X Max only lasts one day. — retained record

> Retained reddit record from reddit:ClaudeCode.

Canonical: https://highsignal.app/data/records/99ef88c088ca48dc

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeCode)  published 2026-10-25 02:35 ingested 2026-10-26 02:39

[original source ↗](https://www.reddit.com/r/ClaudeCode/comments/1xeqs0y/do_i_have_a_workflow_issue_20x_max_only_lasts_one/)

## retained text



```
I posted last time about burning $4000 dollars on the $20 plan. This was partially satire, but I was trying to pick up workflow optimization tips. I did really, though, burn $4000 API credits in one day-- this is apparently the value that the 20X plan gets me weekly.

I don't know how I'm using so much, aside from the workflow itself. I am usually working on one or two projects at once, consistently running, with around 5 subagents per session depending on the task. Constantly grepping and reading unnecessary data doesn't happen in my codebases, since my repos are heavily indexed. Additionally, I have tight constraints/goals, I don't use the advisor, my context window is 200K with auto-compaction, I use very few skills/MCPs, and so forth. My [CLAUDE.md](http://CLAUDE.md) is up-to-date, other state files, I keep them compact.

Is this a problem with the subagent workflow itself? I generally run Opus 5.5 Xhigh for the main model, a handful of subagents, and the rest vary that down to Sonnet 5.5 High.  Maybe there's somethng big I'm missing.
```



document fetched 2026-10-26 00:37 · canonical https://reddit.com/r/ClaudeCode/comments/1xeqs0y/do_i_have_a_workflow_issue_20x_max_only_lasts_one
