AI cost calculatorWhat will your AI cost to run?
Compare monthly running costs across leading models with live prices, then see how much caching, batching and routing can save.
Total usage: 378,400,000 tokens per month
Frontier
Balanced
Efficient
Running open models on your own hardware: cost depends on GPUs and utilisation, not tokens. We can model the break-even for you.
Talk to usAI costs, answered.
How do you estimate monthly AI costs?
Multiply monthly input and output tokens by each model’s price per million tokens. The calculator counts chats, documents and images per user per working day, using editable token assumptions, and prices every model with the same usage.
Where do the model prices come from?
From the LiteLLM price list on GitHub, a community-maintained record of provider prices, checked at least once a day. A dated snapshot is used if the live list is unavailable. Always confirm on the provider’s own pricing page.
How can we reduce our AI running costs?
The biggest levers are prompt caching for repeated context, batch processing for work that can wait, routing simple requests to smaller models, trimming context, and limiting output length. Use the optimisation settings to see their effect.
Is running open models on our own hardware cheaper?
It can be at steady, high volume, or when data must stay in your network. The break-even depends on GPU cost and utilisation rather than tokens. Talk to us and we can model it for your workload.
Ready to audit your enterprise AI readiness?
Discover where AI can make a meaningful difference — and what it takes to get there responsibly.