Vapi Pricing: $0.05 a Minute Plus Models, and What Calls Really Cost

October 15, 2026. Vapi charges $0.05 a minute for hosting, plus the speech-to-text, language model, voice and phone costs of each call at the providers' own prices with no markup. Vapi's own calculator puts 1,000 minutes at $82 to $129 a month with Deepgram, OpenAI and ElevenLabs on its built-in telephony, which works out to about $0.08 to $0.13 a minute. On top of usage you can add a success package: none, Core at $29 a month, Pro at 10 percent of hosting fees with a $999 monthly minimum, or a custom Premier deal. We build AI receptionists and calling agents for businesses, including AI call answering for home services, so this guide shows what a Vapi minute really costs and where the bill grows.

Vapi's prices come from its pricing page, read 15 October 2026. Provider rates inside the per-minute cost change whenever those providers change their prices, so treat the ranges as Vapi's current examples rather than fixed figures.

How a Vapi minute is built

Vapi says every minute has the same parts: speech-to-text, the language model, text-to-speech and transport, each billed at the provider's listed price, with Vapi's $0.05 hosting fee on top. Using the examples on its pricing page:

The language model is the line that moves most: the top of OpenAI's range is nearly six times the bottom. Transport matters for outbound calling. Vapi's own telephony, SIP, WebSockets and Daily WebRTC are free, while carriers charge their own rates: Twilio $0.008 a minute inbound and $0.014 outbound, Vonage $0.00814 and Telnyx $0.0055, according to Vapi's table.

Success packages and what they include

Every package runs on the same platform with unlimited agents, and usage is billed the same way on all of them. Vapi's FAQ notes that the Pro percentage applies to hosting fees only, not model costs. Because hosting is $0.05 a minute, 10 percent of it only exceeds the $999 minimum at roughly 200,000 minutes a month; below that, Pro costs $999.

Add-ons that change the bill

• Extra concurrency: $10 per call line per month. A receptionist that must answer ten calls at once needs either six extra lines on usage-only ($60 a month) or the Core package ($29 a month, ten lines included). Core is cheaper.
• Extra organizations: $20 a month each. An organization is a separate workspace with its own agents, numbers and keys, useful for keeping development and production apart or for agencies running several clients.
• HIPAA: HIPAA-eligible data handling with a DPA and BAA costs $2,000 a month per organization. For a medical or dental practice, that line can be larger than all the call minutes combined.

What real call volumes cost

Using Vapi's example rates of about $0.082 to $0.129 a minute on its own telephony:

• 1,000 inbound calls of 3 minutes (3,000 minutes): about $245 to $387 a month, plus $10 a month for each simultaneous call line beyond four, or $29 for Core, which includes ten.
• The same volume outbound over Twilio, adding $0.014 a minute for transport: about $287 to $429 a month.
• 10,000 minutes a month: about $818 to $1,289 in usage.

Two costs sit outside these figures: the build itself, meaning prompts, integrations with your calendar or CRM and testing, and the compliance work that outbound calling needs. In the United States, the FCC confirmed in a February 2024 declaratory ruling that the TCPA's rules on artificial or prerecorded voices cover AI-generated voices, so outbound programs need consent records and opt-out handling before the first dial; our guide to TCPA-compliant AI calling covers the setup.

How to keep the per-minute cost down

• Pick the model deliberately. In Vapi's own example the language model ranges from under a cent to about 4.5 cents a minute, the widest swing of any line. Test whether a smaller model handles your call script before defaulting to the largest one.
• Keep the prompt and tools lean. Language models bill on the text they read and write, so a long system prompt and a long list of tools are paid for on every turn of every call.
• Use Vapi's own telephony where it fits. Vapi lists its telephony and SIP transport as free, while carrier minutes add up on outbound campaigns.
• Cap call length and handle voicemail. A hard limit on call duration and a voicemail rule on outbound calls stop runaway minutes from a stuck conversation or an answering machine.
• Size concurrency to your real peak. At $10 a line, paying for lines you never use is a quiet monthly cost; missing calls at the busiest hour is a louder one.

Vapi vs Retell on price

The two platforms price the same way, a platform fee plus pass-through models, and end up close. According to our Retell AI pricing breakdown, based on Retell's pricing page as read on 13 October 2026, Retell's infrastructure fee is $0.055 a minute against Vapi's $0.05 hosting, and a standard Retell build comes to about $0.149 a minute. The bigger differences are around the minutes: Retell includes 20 concurrent calls and charges $8 for each extra line, while Vapi includes 4 on usage-only, 10 on Core and charges $10 a line; and Vapi's HIPAA add-on is a flat $2,000 a month per organization. Our Vapi vs Retell comparison covers contracts and features in detail.

Which package to choose

1. Testing or a low-volume line: usage only, with the $5 credits and four concurrent calls.
2. A live receptionist or small outbound program: Core at $29, for ten lines, five numbers, 30-day retention and support with an SLA.
3. Production at scale with uptime commitments: Pro, which costs $999 a month until hosting fees pass about $9,990 a month.
4. Healthcare: budget the $2,000 HIPAA add-on from day one, and compare it with dedicated options in our guide to HIPAA-compliant AI receptionists.

For a trades business, the cheapest Vapi build is rarely the best one; the agent has to book jobs and route emergencies properly, which is what our AI answering service guide for HVAC companies walks through. If you would rather not assemble it yourself, our AI calling team builds and runs these agents.