#1 · The Claude Model Family

One model family, three speeds

The right model for the right job

Haiku

Lightning fast

Output

95–150 tok/s

Cost / MTok

$1 in · $5 out

Context

200k tokens

Best for

High-volume tasks

Sonnet

Balanced

Output

~53 tok/s

Cost / MTok

$3 in · $15 out

Context

1M tokens

Best for

Day-to-day work

Opus

Deep thinker

Output

~45 tok/s

Cost / MTok

$5 in · $25 out

Context

1M tokens

Best for

Complex reasoning

1 / 1