Price with checked integers · AI Engineering

Stage 3 of 4. Rust. Plan about 2 hours.

The fixture rate card uses nano-dollars per token. Charge uncached input, cached input and output at separate rates. Every multiplication and sum is checked for overflow, and no floating-point rounding can silently change the bill.

pj token counter and cost meter 3

Implementation boundary

rust

pub fn cost(usage:&Usage,rates:&Rates)->Result<u64,Error>

Primary reference: Official reference.

Worked Orchard case

Before coding, review Rust ownership and Result and Tokenizers. Complete stage 2 first.

Use supplied integer rate cards in nano-dollars per token. With input=2, output=5 and cached=1, the recorded request costs 260 nano-dollars. These fixture rates are not current provider prices.

text

60 uncached * 2 = 120
40 cached * 1 = 40
20 output * 5 = 100
total=260 nano_dollars

Build and inspect

Multiply and add with checked u64 operations. Do not round each component through floating-point dollars.

Implement the stage in your learner workspace. The CLI helpers are provided adapters and import your functions; they do not substitute the reference solution.

bash

python3 scripts/project_test.py token-counter-and-cost-meter --stage 3 --path learning-artifacts/token-counter-and-cost-meter

Predict the intermediate state above, then run the stage. A fresh stub fails; a passing reference run does not establish completion of your learner workspace.

Investigate next

How would an overflow affect budget admission if arithmetic wrapped silently?