Prices at a glance
On November 6, 2026, published AI voice agent prices fell into three models. Developer platforms charge per minute, from $0.05 a minute for hosting on Vapi to $0.07 to $0.31 on Retell and $0.14 on Bland, usually before you add the language model and telephony. Managed answering services charge a monthly plan with included minutes: $29 to $449 to start, with Botcadence's Voice AI plan at $299 for 1,500 minutes and $35 per 200 more. Per-call services such as Smith.ai charge $1.67 to $3.00 a call. At 1,500 minutes a month, 500 three-minute calls, the spread runs from $75 for Vapi's hosting alone to $2,344 on a live receptionist plan. The calculator below prices your own volume.
We build Botcadence, so read our row the way you would read any vendor's. Every figure links to the page it came from.
The three pricing models
| Model | How you pay | Who uses it | What raises the bill |
|---|
| Per-minute platform | A rate per connected minute, plus the language model and telephony where they are not included | Retell, Vapi, Bland, ElevenLabs | Every minute, and the add-ons you switch on |
| Monthly plan with included minutes | A flat fee for a block of minutes, then an overage rate or a bigger plan | Botcadence, Dialzara, My AI Front Desk, Rosie, Sameday | Passing the included minutes |
| Per call or per caller | A plan with included calls, then a rate per extra call or unique caller | Smith.ai, Goodcall | Every extra call; spam rules decide what counts |
Live receptionist services bill per minute of operator time, from $1.49 to $2.99, and are the baseline every AI vendor compares itself to.
Prices by vendor, checked November 6, 2026
| Vendor | Model | Price | Included | Extra | What is not included |
|---|
| Vapi | Platform | $0 to start; Core $29 a month; Pro from $999 | Nothing; $5 credits | $0.05 a minute hosting | Models at cost; Twilio $0.008 to $0.014 a minute; $10 per extra concurrent line |
| Retell AI | Platform | Pay as you go | Nothing; $10 credits | $0.07 to $0.31 a minute | Telephony $0.015 a minute; knowledge base $0.005 a minute |
| Bland AI | Platform, models included | $0 (Start) or $299 a month (Build) | Nothing | $0.14 a minute (Start), $0.12 (Build) | Transfer time $0.04 to $0.05 a minute |
| ElevenLabs Agents | Platform | Free, then $6 to $990 a month across five plans | 15 to 12,375 minutes | $0.08 a minute | The language model, billed by usage |
| Botcadence | Plan | $299 a month | 1,500 minutes | $35 per 200 minutes | Nothing for voice; Meta's WhatsApp fees if you use that channel |
| Dialzara | Plan | $29 to $349 a month | 60 to 1,000 minutes | 35 to 48 cents a minute | Done-for-you setup from $1,500 a month |
| My AI Front Desk | Plan | $99 a month ($79 yearly) | 200 minutes | 25 cents a minute | Nothing stated |
| Sameday | Plan | $449 or $789 a month | 500 or 1,000 minutes | Not published | API on Enterprise only |
| Smith.ai AI receptionist | Per call | Free for 25 calls, then $150 or $500 a month | 25 to 300 calls | $1.67 to $3.00 a call | Nothing stated |
| PATLive (live) | Per minute | $49 to $6,899 a month | 0 to 5,000 minutes | $1.49 to $2.99 a minute | Human-only; no AI |
Synthflow publishes no prices; it sells annual enterprise agreements. Goodcall prices per unique caller ($79 for 100), which doesn't convert to minutes; see the best AI answering services list.
What 500, 1,500 and 3,000 minutes cost
Three-minute calls, so 167, 500 and 1,000 calls. Platforms are shown at their published rate without the language model or telephony unless the vendor includes them; add $0.02 to $0.10 a minute for those on Vapi and Retell.
| Vendor | 500 minutes | 1,500 minutes | 3,000 minutes |
|---|
| Vapi (hosting only) | $25 | $75 | $150 |
| Bland Start | $70 | $210 | $420 |
| Retell (at $0.15 plus telephony) | $83 | $248 | $495 |
| ElevenLabs (at $0.08 a minute, plan fee and model excluded) | $40 | $120 | $240 |
| Botcadence Voice AI | $299 | $299 | $579 |
| My AI Front Desk | $174 | $424 | $799 |
| Dialzara | $199 | $524 | $1,049 |
| Sameday | $449 | $789 | Not published |
| Smith.ai AI receptionist | $334 | $835 | $1,670 |
| PATLive (live) | $946 | $2,344 | $4,879 |
Two things stand out. Platforms are cheapest on paper and most expensive in time: someone has to build and maintain the agent, the prompts, the calendar connection and the transfer logic, which is why small businesses usually buy them through an agency at $500 to $3,000 a month. Managed plans cross over at different volumes: Botcadence costs more than ElevenLabs at every row and less than Dialzara above about 700 minutes.
What else to price
- Billing increments. Botcadence bills a 15-second first step, then 30-second steps. Live services and some AI plans round every call up to the minute, so a 10-second spam call costs a full minute.
- Concurrency. Vapi charges $10 a month per extra simultaneous line; Bland gives 10 on Start. Every service has a limit; ask for it in writing.
- Transfers. Bland bills transfer time at $0.04 to $0.05 a minute. On Botcadence the escalation calls that wake your on-call tech use minutes.
- Numbers and telephony. Retell and Vapi charge for carrier minutes on top; ElevenLabs and Botcadence do not.
- Compliance add-ons. Vapi's HIPAA add-on is $2,000 a month. Botcadence is not HIPAA compliant and does not sell to practices that need a BAA.
- Setup. Self-serve plans carry no setup fee on Botcadence; Dialzara's done-for-you tier starts at $1,500 a month.
Is it worth it against a person?
A 1,000-minute month costs $1,499 on PATLive and $299 on Botcadence; the $1,200 difference pays for the escalation and chat channels many times over. But the comparison that matters is cost per booked job, not cost per minute. Track answer rate and booking rate before and after; the HVAC booking rate guide shows the formulas and the AI answering service cost guide covers the hidden fees. To hear what the minutes buy, call the AI voice agent demo line, or read the AI receptionist and inbound voice pages for what a managed plan includes.