Anthropic API

Claude models, with prompt caching priced separately and long-context tiers billed at a different rate.

Official docs ›

Price per million tokens

List price, read from the provider's own page · verified 2026-08-04

ModelInput / 1MOutput / 1MCached in
claude-fable-5$10$50
claude-haiku-4-5$1$5
claude-opus-5$5$25
claude-sonnet-5$2$10

Compared with

Missing a model, or a price that moved? Prices here are only ever copied from the vendor's own page and stamped with the date they were read — never estimated. A verified run measures what the price table cannot: time to first token, throughput under concurrency, function-calling reliability.