Is Retell AI or Vapi better?
Retell AI is better when you want one bill and a quick start: its $0.07 to $0.31 a minute covers the voice engine, the voice and the model, telephony adds $0.015 unless you bring your own SIP, and a concurrency limit of 20 costs nothing. Vapi is better when you want to pick and pay each provider yourself: it charges $0.05 a minute to host the call and passes speech-to-text, model and voice through at provider rates, which Vapi estimates at $82 to $129 per 1,000 minutes. Both are developer platforms, and both prices were read on November 13, 2026.
Key takeaways
- Retell's default setup comes to about $0.11 a minute, or $0.125 with its telephony; Vapi's own estimate is about $0.08 to $0.13.
- Retell includes a concurrency limit of 20 and charges $8 a month for each extra. Vapi's limit is 4 on the free plan and 10 on the $29 plan, with extras at $10 a month each.
- Retell lists a health-data agreement under Enterprise. Vapi sells it as a $2,000-a-month add-on.
- Retell bills per second and does not bill failed calls; silence and hold time are billed.
- Neither price includes the work of building and maintaining the agent.
Retell AI vs Vapi at a glance
| Retell AI | Vapi |
|---|
| Pricing model | Pay as you go, all parts itemized | Hosting fee plus provider pass-through |
| Published rate | $0.07 to $0.31 a minute | $0.05 a minute hosting; $82 to $129 per 1,000 minutes estimated |
| Monthly plans | None required; Enterprise custom | $0 / $29 / Pro with a $999 minimum / Premier custom |
| Free credit | $10 | $5 |
| Telephony | $0.015 a minute through Twilio or Telnyx; none for your own SIP | Vapi telephony and SIP free; Twilio $0.008 inbound and $0.014 outbound; Telnyx $0.0055 |
| Phone numbers | $2 a month | 1 on the free plan, 5 on the $29 plan |
| Concurrency | Limit of 20 included; $8 a month each after | Limit of 4, 10 or 30 by plan; $10 a month each after |
| Health-data agreement | Enterprise | $2,000 a month |
| Add-ons | Knowledge base $0.005 a minute; removing personal data $0.01; branded calls $0.10 each | Not itemized on the pricing page |
What 1,000 and 5,000 minutes cost
| Minutes a month | Retell AI, default setup | Retell AI, full range | Vapi, own estimate |
|---|
| 1,000 | $110 | $70 to $310 | $82 to $129 |
| 5,000 | $550 | $350 to $1,550 | $410 to $645 |
Vapi's estimate scales in a straight line because no plan includes minutes; add $29 a month if you need the $29 plan's numbers and concurrency. Retell's default figure is from its own calculator and is before telephony, which adds $15 per 1,000 minutes on Retell's numbers; Vapi's own telephony has no transport fee.
When to choose Retell AI
- You want telephony, voice and model on one invoice.
- You expect bursts: a concurrency limit of 20 is included.
- You want per-second billing and no charge for failed calls.
The parts of the rate are in Retell AI pricing, and Botcadence vs Retell compares it with a finished agent.
When to choose Vapi
- You already have contracts or credits with speech and model providers.
- You want to swap providers per assistant without changing platform.
- You need a published uptime commitment, which starts on the Pro plan.
See Botcadence vs Vapi.
Where Botcadence differs from both
Retell and Vapi are platforms: you or an agency build the agent. Botcadence, which we build, is a finished AI voice agent at $299 a month for 1,500 minutes, with booking, warm transfers, on-call alerts and chat channels included, set up for service businesses such as HVAC and restoration companies. At about 20 cents a minute when all 1,500 minutes are used, it costs more than these platforms' typical published rates. It offers no health-data agreement, has no SOC 2 report, and a new account answers one call at a time until the limit is raised. For the wider field see best AI voice agents and the AI voice agent pricing comparison.