Question 1

How is inference cost calculated?

Accepted Answer

For API models, cost = (input tokens * input price + output tokens * output price) * requests. For self-hosted, cost = GPU hourly rate / max throughput (requests per hour).

Question 2

What is break-even?

Accepted Answer

Break-even shows the number of requests per day where self-hosted total cost equals API cost. Above that, self-hosted may be cheaper.

⚡ AI Inference Cost Calculator

📊 Cost summary

📋 All options comparison