Model catalog

Usage cost calculator

Plug in how many million tokens you expect, pick a model (or enter rates), and see list-rate dollars. Assumptions are yours. This is not a vendor bill.

Models · Plans · Install CLI to measure real sessions

Model

Input MTok

Output MTok

Cache read MTok

Cache write MTok

Using cheapest list rates for Llama 3.1 405B Instruct ($10.68 in / $10.68 out per MTok).

INPUT

$106.80

10 MTok × $10.68

OUTPUT

$21.36

2 MTok × $10.68

CACHE READ

$42.72

40 MTok × $1.07

CACHE WRITE

$0.00

2 MTok × $0

List-rate total for this shape

$170.88

Million tokens × published $/MTok. Cache rates use the cheapest listed cache field when the model is known; otherwise the custom rates you entered. This is not a bill and not a subscription comparison.

Real usage, not a guess

The CLI reads the session logs your harnesses already write and prices them the same way.

$ curl -fsSL https://tokenstat.ai/install.sh | bash