grok-4.20

Pricing

Input

from $1.25per 1M tokens

Output

from $2.50per 1M tokens

Cache read

from $0.20per 1M tokens

A price marked "from" is where the model starts: what a call costs depends on which provider serves it and, for some models, the time of day.

Long context, a prompt above 200,000 tokens: $2.50 input, $0.40 cached input, $5 output per 1M tokens.

Billing

On our keys, usage of this model is billed at exactly what the model's provider charges for each call, with no markup. On your own keys, your provider bills you directly at its own rates.

Price updated October 15, 2026.