Cartesia cuts call rates. Check the full voice-AI budget.

Rebuild the call budget

If your team is considering AI for customer calls, a lower minute rate is a reason to revisit the budget. Start with the full cost of completing a useful request, including the time people spend checking the result.

Cartesia now lists its Managed Agents calling rate at $0.03 per minute, beside the previous $0.06 rate. Managed Agents runs live voice conversations for a business. Pricing trackers recorded the change by November 2; the exact effective date for an existing contract needs checking.

The lower public rate covers the Free, Pro, Startup and Scale plans. Enterprise pricing is custom. It does not mean every part of the call costs half as much.

Read the announcement: Cartesia: current plans, previous and new call rates, and separate charges.

Separate the three charges

Cartesia lists a further $0.014 per minute when you use a phone number it provides. The AI model, meaning the system that interprets the conversation and chooses a reply, is billed separately according to the model and text it processes.

For 1,000 minutes, the base calling charge is $30. A Cartesia-provided number adds $14. Those two components total $44 before AI-model usage and before working through subscription allowances. This is an illustration, not a final invoice.

The published plans include prepaid dollars for agent calls alongside separate credits for speech products. Ask how the included balance and extra usage affect what you actually pay. Commercial-use licensing is listed from the $5-a-month Pro plan.

Check the model and existing bill

Pricing trackers also report that a limited free allowance for AI-model use on agents created in Cartesia’s interface has ended. The current primary pages confirm separate billing, but do not establish every existing customer’s transition date. Check your account before assuming the lower calling rate reduces the whole bill.

Have the person building the call flow show which model is selected and its current price. Changing that choice can change the cost and the replies, so review both together. Keep your existing contract in the comparison.

Our perspective: the useful measure is cost per correctly completed customer request. A cheaper call that needs substantial correction or a second conversation may not be the better service.

Give the trial two owners

The budget owner can check the bill while the service owner checks the experience. Use approved test information for one repeated request before expanding live calls. Agree which actions need a person’s approval and how a customer reaches someone when the AI cannot help.

A lower running cost may make a previously expensive task worth reconsidering. Let the measured result decide how far to take it. We have not tested the service.

1. Before the next renewal or expansion, confirm the rate and model charges that apply to your account.
2. Price a representative call after plan allowances, then include setup, review and ongoing support.
3. Check completed requests, correction time and the handover to a person.
4. Confirm the plan can handle the number of calls expected at once before approving wider use.