Pricing

What the models cost, plus 5%.

That is the only fee. No subscription. No charge per AI teammate. No equity. Hosting at the cloud’s own price. Every saving Maestro finds goes to you.

Apply for access

$10.00model usage $0.50the 5% fee $10.50total

Three lines on the bill

Models.
Each model at its list price, plus 5%. Every task shows its model and its cost in the record.
Sponsored models.
Free. AI labs and cloud providers sponsor models through the Zero Build Foundation. Work on a sponsored model costs you nothing.
Foundation
Hosting.
At cost. ZB Cloud passes each cloud’s own price through, with nothing added.

Why it costs less

Same task. 100 times the price.

The best model for a task is rarely the priciest one. Maestro routes each task to the cheapest model that does it well, and only a failed check pays the top price.

Price per 1 million output tokens

$50.00Claude Fable 5.1Specs
$20.00Claude Opus 5.5Review
$10.00GPT-6.1 SolPlanning
$10.00Claude Sonnet 5.5Tests
$6.00Grok 4.7Research
$3.96DeepSeek-V4-ProCode
$0.60Mistral Small 4Tagging
$0.47Qwen3.8-FlashExtraction

Fable to Qwen: 106 times.

Published list prices per 1 million output tokens, standard tier, 10 October 2026. Full table on the developers page.

What a build costs

Same store. Same work. 96% less.

One model for everything$720.00

Routed by Maestro$29.10

20 systems
3,559 tasks
12 models

Example build, not a customer result. One model = Claude Fable 5.1 at list price on 10 October 2026. The $29.10 includes the 5% fee. Hosting is separate.

See the whole build

One balance

One balance. Every model. Every cloud.

Add to your balance once. Every model and every cloud draws from it. The record shows every task, its model, its cloud and its cost, so the bill is never a surprise.

Three questions

Bring an idea.