claude tokens

per-model api cost and subscription-budget equivalents

0
characters
0
words
0
input tokens
1.00× 0 out
modelinputoutputtotalper 1k runs

pricing source

api rates: anthropic published pricing per million tokens (input / output). subscriptions: anthropic.com plans.

token estimate

claude uses a bpe tokenizer (~3.5 chars / token for english prose). actual token count can vary ±10% from this heuristic — non-ascii, code, and structured data tokenize differently.

subscription equivalents

"runs per month" = monthly plan price ÷ this prompt's sonnet 4.6 total cost. it shows how many of this exact prompt the subscription's dollar value would cover if billed at api rates. real subscriptions have rate limits, not token caps — but this gives you a sense of the effective subsidy.

example: at 10× api-equivalent usage, pro is paying for itself; at 25×, max 20x is paying for itself.