Spend ledger
The honest numbers.
Every attempt counts — including the work that never shipped.
Cost / approved asset
$0.015833
Across 9 accounted runs, including discarded attempts.
Spend prevented
$0.015000
Policy-blocked work · zero provider calls were made.
Waste ratio
0.0%
6 of 9 runs produced a published asset.
Live ledger unavailable · verified B2 snapshot from 29 Jul 2026
Spend by model
| Model | Provider | Runs | Total (USD) | Mean (USD) |
|---|---|---|---|---|
| gpt-image-2 | openai | 9 | $0.095000 | $0.013571 |
Spend by project
| Project | Runs | Approved assets | Total (USD) |
|---|---|---|---|
| prj_t24_proof | 2 | 2 | — |
| prj_dara_smoke | 2 | 2 | — |
| prj_dara_ledger_smoke | 1 | 1 | — |
| prj_dara_qa_smoke | 2 | 1 | — |
| prj_policy_proof | 1 | 0 | — |
| prj_recovery_probe | 1 | 0 | — |
Spend by month
| Month | Runs | Total (USD) |
|---|---|---|
| 2026-07 | 9 | $0.095000 |
Null cells are values the recorded evidence did not preserve; they are shown as unknown rather than filled in. Conservative low-quality policy reservations are used where a provider did not return settled cost.