AI Cost Calculator

Estimate model cost per accepted job.

Compare your current model with one model from Acta's live catalog. Every input is visible, and the result remains an estimate rather than observed savings.

Enter your assumptions

Enter rates in USD per token, not per million tokens. Candidate rates load from the selected model and remain editable.

Retry rate is the estimated share of requests repeated once. Accepted-job rate is the estimated share of monthly requests that produce a result your business accepts.

Limits: 100 million monthly requests, 1 million input or output tokens per request, $1 per token, and 100% for retry and accepted-job rates.

Shared Workload assumptions

Monthly requests Average input tokens per request Average output tokens per request

Current assumptions

Current input rate (USD per token) Current output rate (USD per token) Current retry rate (%) Current accepted-job rate (%)

Candidate assumptions

Candidate active Acta LLM Candidate input rate (USD per token) Candidate output rate (USD per token) Candidate retry rate (%) Candidate accepted-job rate (%)

Inputs stay in the GET URL for sharing. Acta does not add those values to first-party product events.

Do not submit prompts, responses, credentials, invoices, or personal data. This worksheet needs only aggregate numeric assumptions and an active catalog model.

Test quality before changing production.

Apply for a design partner assessment