TokenWatch — Compare LLM API Costs

PROVIDERS

Search by provider

Search by model

USAGE

Your Usage

Compute by:

Total tokens Million tokens (e.g. 1000 = 1B tokens)

Input % Tokens sent to the model

Cached input % Cached prompt tokens (discounted)

Output % Tokens generated by the model

Input: 25M · Cached: 970M · Output: 5M 100%

Presets:

Advanced: cache write

Cache write tokens (M, one-time) Amortize over N requests

PRICING

Group by: Show Org

#	Org	Provider	Model	Input $/M	Output $/M	Cache $/M	Context	Speed	Total Cost
Select a model or provider to see pricing.