Side-by-side cost instrument

Compare AI model costs

Run one workload through up to four models. TokenMeter applies each provider's current cache and context-band rules before ranking the result.

Comparison workloadUp to 4 models

Set your workload

1/4 models

Workload presetChoose an example starting point or enter your measured workload directly.

Monthly input tokensMonthly output tokensCached input tokensAverage input tokens per requestUsed only where providers charge a long-context rate.

calibrated

Google

Lowest cost

$26.00/ month

Rate alternativessource-backed only

Cache-readSet cached input tokens above zero to estimate cache-read savings.

Unavailable

BatchPublished batch rate unavailable for one or more required dimensions.

Unavailable