Vapi vs Retell pricing (2026)
Both charge a per-minute platform fee and then bill models and phone minutes. The difference is how: Vapi passes providers through at cost on top of $0.05/min; Retell charges $0.055/min voice infrastructure and lists its own per-minute price for each model and voice.
Bottom line at our default assumptions
- Same premium-ish stack (Deepgram STT, GPT-5 mini, ElevenLabs Flash, Twilio): Vapi $0.0858/min vs Retell $0.1180/min.
- Budget stack: Vapi $0.0680/min (Deepgram Aura-1 voice, Telnyx SIP) vs Retell $0.0812/min (Retell platform voice, BYO SIP).
- The biggest swing is TTS: Retell prices ElevenLabs Flash at $0.04 per call minute, while ElevenLabs' own API price works out to about $0.02 per call minute if your agent talks half the time. Retell's cheaper house voices ($0.015/min) close the gap.
These are estimates from list prices. Your LLM cost depends heavily on prompt length; edit the assumptions in the calculator.
| Layer | Component | $/call-min | Basis |
|---|---|---|---|
| Platform | Vapi Hosting | $0.0500 | source |
| STT | Deepgram Nova-3 Monolingual, streaming | $0.0048 | Passed through at cost. source |
| LLM | OpenAI gpt-5-mini | $0.0025 | Token price x our assumptions. source |
| TTS | ElevenLabs Flash / Turbo | $0.0200 | $/1K chars x 1,000 chars/min x 50% talk. source |
| Telephony | Twilio Local number, inbound | $0.0085 | Your own Twilio number. source |
| Total | $0.0858 | $85.78 per 1,000 min | |
| Layer | Component | $/call-min | Basis |
|---|---|---|---|
| Platform (incl. STT) | Retell AI Voice Infra | $0.0550 | source |
| LLM | Retell: GPT 5 mini | $0.0080 | Retell list price per minute source |
| TTS | Retell: ElevenLabs Flash | $0.0400 | Retell list price per minute source |
| Telephony | Retell: Retell Twilio (US) | $0.0150 | Retell list price per minute source |
| Total | $0.1180 | $118.00 per 1,000 min | |
| Layer | Component | $/call-min | Basis |
|---|---|---|---|
| Platform | Vapi Hosting | $0.0500 | source |
| STT | Deepgram Nova-3 Monolingual, streaming | $0.0048 | source |
| LLM | OpenAI gpt-5-mini | $0.0025 | source |
| TTS | Deepgram Aura-1 | $0.0075 | source |
| Telephony | Telnyx Elastic SIP, local inbound | $0.0032 | BYO Telnyx SIP trunk. source |
| Total | $0.0680 | $67.98 per 1,000 min | |
| Layer | Component | $/call-min | Basis |
|---|---|---|---|
| Platform (incl. STT) | Retell AI Voice Infra | $0.0550 | source |
| LLM | Retell: GPT 5 mini | $0.0080 | Retell list price per minute source |
| TTS | Retell: Retell platform / Cartesia / OpenAI / Inworld / Minimax / Fish voices | $0.0150 | Retell list price per minute source |
| Telephony | Retell: Custom SIP (BYO carrier) | $0.0000 | Retell charges $0; your carrier bills separately. source |
| Carrier | Telnyx Elastic SIP, local inbound | $0.0032 | Telnyx inbound SIP. source |
| Total | $0.0812 | $81.20 per 1,000 min | |
Fees beyond the per-minute rate
| Vapi | Retell | |
|---|---|---|
| Platform fee | $0.05/min hosting, prepaid | $0.055/min voice infra |
| Model pricing | Provider rates, no markup (per Vapi) | Retell's per-minute list (e.g. GPT 5 mini $0.008, Gemini 3.8 Flash $0.048) |
| Phone numbers | Billed by your telephony provider; free tier includes 1 number | $2/mo per number; Twilio US minutes $0.015; custom SIP free |
| Concurrency | Free tier: 4 concurrent calls | First 20 concurrent free, then $8/concurrency/mo |
| Add-ons | See vendor | Knowledge base +$0.005/min; PII removal +$0.01/min |
| Enterprise | Pro: 10% of hosting fee, $999/mo minimum | Contact sales |
Sources: Vapi pricing, Retell pricing, checked 2026-10-17. Retell STT-included is our inference from its published $0.07/min floor.