0% 5h quota used · resets in 2h live · 08/09/2026 12:04 CEST

The books

Every run priced from #1. Ceilings are per-run and per-day per country, with $540 a day as the fleet backstop.

API-equivalent $4,536
across 516 runs · what this would have cost on the API
Cash spent $0.53
real money on prepaid providers · $0 on the subscription
Agent time 98h 36m
13,170 turns
Days over the ceiling 0
ceiling $540 a day
Runs over the ceiling 0
each against its own country's per-run ceiling

Spend per day

red where the day broke $540 — the fleet backstop, the sum of every country's daily ceiling
daily ceiling $540 2026-07-02: $10.692026-07-03: $1.452026-07-09: $02026-07-11: $0.552026-07-14: $0.762026-07-15: $4.172026-07-16: $0.392026-07-17: $0.372026-07-22: $1402026-07-23: $13.232026-07-24: $82.622026-07-25: $1072026-07-26: $1.262026-07-27: $6.532026-07-28: $0.422026-07-29: $6.002026-07-30: $15.892026-07-31: $1.582026-08-04: $20.522026-08-05: $16.592026-08-08: $22.442026-08-09: $0.782026-08-10: $9.282026-08-11: $12.662026-08-12: $23.372026-08-13: $10.902026-08-14: $6.692026-08-15: $2.092026-08-16: $1.882026-08-17: $10.322026-08-18: $35.312026-08-19: $1422026-08-20: $1972026-08-21: $1532026-08-22: $23.112026-08-23: $1.962026-08-24: $58.962026-08-25: $1172026-08-26: $1052026-08-27: $2902026-08-28: $3302026-08-29: $3782026-08-30: $2192026-08-31: $4242026-09-01: $2662026-09-02: $2522026-09-03: $1942026-09-04: $2312026-09-05: $1532026-09-06: $1832026-09-07: $1992026-09-08: $53.34 07-0207-1607-2608-0408-1208-1808-2408-3009-05

What the money went on

by task, all of the selection
de:wasserversorger-enrich $1,891 63 runs
_global:sentry-triage $32.71 7 runs
tn:seo-audit $15.92 13 runs
_global:pr-review $15.04 22 runs
tn:seo-fixes $13.75 5 runs
fr:sentry-triage $12.87 4 runs
tn:content-freshness $11.11 3 runs
fr:pr-review $7.42 17 runs
tn:content $4.50 1 runs
_global:monthly-report $3.21 2 runs
tn:business-review $1.57 2 runs
tn:ctr-optimizer $1.56 1 runs
tn:decisions-answer $0.91 2 runs
_global:gate-probe $0.72 3 runs
_os:dispatch $0.38 19 runs

By workspace

de $4,411 369 runs
_global $54.65 59 runs
tn $49.32 48 runs
fr $20.29 21 runs
_os $0.38 19 runs

By model

script runs cost nothing by design
sonnet $4,524 421 runs
opus $10.85 6 runs
haiku $1.10 22 runs
script (no model) $0 67 runs

Waiting for a slot

the cost of the concurrency cap, in time rather than money — measured on 140 agent runs since it started being recorded
Runs that waited 13
of 140 measured · 9%
Time spent queued 5h 03m
work that was ready and could not start
Longest wait 36m
TaskRuns that waitedTotalWorst
de:netzbetreiber-replies104h 54m36m
_global:monthly-report15m5m
de:wasserversorger-enrich12m2m
_global:pr-review11m1m

Month by month

MonthRunsFailedTurnsAgent timeCostPer run
2026-0910432,94428h 08m$1,531$14.72
2026-08262128,65160h 16m$2,612$9.97
2026-07150471,57510h 11m$393$2.62

Runs that cost far more than their task usually does

over 3× the task median, and over a dollar

Nothing here was actually invoiced: runs go through the Claude subscription, and cost_usd is what the same tokens would have cost on the API. That is the point — 71 of these 516 runs are deterministic scripts and cost nothing at all, and the gap between those two numbers is the argument for moving work out of agent turns.