Spend report · @stealth-promptsmith-40

$1,967/mo burn

ranked against 38 devs · median $473/mo

4.2×
Burns 4.2× the median dev — $1,967/mo vs $473. Top 16% of the board.
86%
86% of tokens re-read context already loaded; only 2.95% was newly written code. Paying to reload, not to create.
84%
Opus drives 84% of the bill — the single biggest cost lever on this profile.
Where this landspercentile vs cohort · ◆ median
Monthly spend$1,96784th pctl
Tokens that were actual code2.95%79th pctl
Spent re-reading context86%24th pctl
Opus reliance84%66th pctl
Burn per active day$6684th pctl
Model mixby cost
Opus 84.2%Everything else 15.8%
Cut the burncomputed from these numbers
Route routine work to Sonnet — saves ~$397/mo
Keep Opus for architecture and hard debugging; /model sonnet for edits, tests, boilerplate.
Cut context re-reads — saves ~$236/mo
/compact at natural breakpoints, /clear between tasks, trim CLAUDE.md — every turn re-reads it.
Cohort distributionprojected / month
<$100 5
$100–300 11
$300–600 7
$600–1k 5
$1k–2k you
$2k–5k 3
$5k+ 2
$ npx tokenrot
Out-burns 84% of developers. See where you land.

The Burn Report

Once a month: what the cohort actually spent, which models got throttled, and where the money leaked. Written from the board's data — no spam, unsubscribe anytime.