One prompt. No frameworks. Every visual drawn in code.
Estimated at standard API rates
| Uncached input996,714 tokens × $10 / 1M | $9.97 |
| Cached input29,180,032 tokens × $1 / 1M | $29.18 |
| Output + reasoning434,931 tokens × $50 / 1M | $21.75 |
| 22m 34sCoordinator + all 10 builders | 1 prompt |
| Token category | Tokens | Rate / 1M | Estimate |
|---|---|---|---|
| Uncached input | 996,714 | $10.00 | $9.97 |
| Cached input | 29,180,032 | $1.00 | $29.18 |
| Cache writes reported | 0 | $12.50 | $0.00 |
| Output, including reasoning | 434,931 | $50.00 | $21.75 |
| Total | 30,611,677 | $60.89 |
Line items are rounded separately; the total uses unrounded values. The logs report zero cache-write tokens.
Sum final cumulative usage once for the coordinator and each of 10 builders. Builder follow-ups are already included. Reasoning is included in output, not added twice. No request exceeded 272K input tokens. Logs report zero cache-write tokens. Standard rates; no Fast-mode multiplier. Run used Codex on a Pro plan.
This estimate applies OpenAI’s standard API prices to the recorded token usage. It does not measure the charge against the Pro subscription or apply a Fast-mode premium. Pricing checked September 6, 2026.
Includes planning, all ten builders, coordinator checks, builder follow-up fixes and the original gallery. Excludes later hosting and promotion work.
The original 100 HTML pages are preserved in the public gallery. Generated page count is not an independent quality score.
