Independent buyer reference. Not affiliated with Gong, Clari, ZoomInfo, 11x, Artisan, Regie.ai, Vapi, Retell, Bland, or any AI sales vendor. Prices verified May 2026; confirm before purchase. Legal overview | FAQ
Voice AI InfrastructurePublic pricing; bundled rate

Retell AI Pricing in 2026: The $0.07 Bundle and Why It Often Beats Vapi

Retell publishes a $0.07 per minute bundled rate that includes STT, LLM, TTS, and the platform layer. Add telephony and you land at $0.085 to $0.20 per minute all-in. For typical sales deployments this beats Vapi's component-stacked cost by a meaningful margin. Here is the breakdown for the 2026 procurement cycle.

Last verified May 2026. Retell pricing changes periodically; always confirm before purchase.

$0.07/min

Bundled (platform + STT + LLM + TTS)

$0.014-$0.03/min

Telephony pass-through

$0.085-$0.13/min

All-in typical (mid-stack)

$0.04-$0.05/min

Enterprise (500K+ min/mo)

§What the Bundled $0.07 Per Minute Actually Includes

Retell's strategic bet is that most developers building voice agents do not want to assemble five providers; they want a single per-minute number for the whole stack and to focus on prompt engineering, business logic, and CRM integration. The $0.07 bundle is what that bet looks like at retail.

Included in $0.07/minute bundled rate

  • + Retell platform: orchestration, VAD, interruption handling, function-calling, recording delivery
  • + STT: Deepgram Nova-2 by default; AssemblyAI as alternative at no per-minute uplift
  • + LLM: GPT-4o-mini, Claude Haiku, Gemini Flash, or equivalent default tier
  • + TTS: ElevenLabs Turbo v2.5, PlayHT, or equivalent default tier
  • + Call recording and transcription to your destination (Webhook, S3, GCS)
  • + Knowledge base ingestion and retrieval (RAG) up to 100MB per agent
  • + Function calling and tool use (CRM lookups, calendar booking)
  • + Multi-language out of the box (40+ languages on default TTS, varies by voice choice)

Adds uplift beyond $0.07/minute

  • + Premium LLM (GPT-4o full, Claude Sonnet 3.5/3.6): $0.06 to $0.12 per minute uplift
  • + Premium TTS (ElevenLabs Multilingual v2, custom-cloned voices): $0.06 to $0.10 per minute uplift
  • + Telephony pass-through (Twilio, Telnyx, Vonage): $0.014 to $0.03 per minute outbound, $0.0085 inbound
  • + Phone number rental (US local): $1 to $2 per month per number, plus carrier A2P 10DLC registration fees
  • + Knowledge base above 100MB: tiered enterprise pricing
  • + Concurrent call capacity above 100 concurrent: enterprise reserved capacity tier

§All-In Cost: Three Reference Sales Stacks

For sales deployments, the realistic 2026 per-minute all-in cost on Retell sits in a tight band of $0.085 to $0.20 depending on LLM and TTS choice. The bundled rate compresses the per-component variability that dominates Vapi's stack.

ComponentBudgetMidPremium
Retell bundle$0.07$0.07$0.07
LLM uplift$0 (default)$0.04 (GPT-4o-mini upgrade)$0.12 (GPT-4o full)
TTS uplift$0 (default)$0.04 (ElevenLabs Turbo)$0.10 (ElevenLabs Multi v2)
TelephonyTelnyx $0.005Twilio $0.014Twilio $0.014
Per-minute total$0.075$0.164$0.304
Per 5-min call$0.38$0.82$1.52

The Retell-vs-Vapi math: The mid stack on Retell costs $0.164 per minute. The equivalent mid stack on Vapi costs $0.208 per minute. The $0.044 per-minute gap is around 21 percent savings on Retell at the mid tier. At 50,000 minutes per month that is $2,200 per month, or $26,400 per year. The savings narrow at the budget tier and at the premium tier where component pricing dominates the bundle benefit.

§Monthly Volume Math for Sales Teams

ScenarioMonthly minRetell midVapi midSavings
Pilot: 3K min3,000$492$624$132 (21%)
Mid: 15K min15,000$2,460$3,120$660 (21%)
High: 50K min50,000$8,200$10,400$2,200 (21%)
Scale: 250K min (enterprise tier kicks in)250,000$32K-$38K$45K-$55K$13K-$17K (28-31%)

The savings percentage holds steady at around 20 percent at standard volume and widens to nearly 30 percent at enterprise volume where Retell's bundle negotiation with STT and LLM providers compounds. The flip side: Retell offers less component flexibility (BYOK is limited; custom STT routing is harder; custom telephony pipelines beyond Twilio and Telnyx require enterprise contact). Teams that need fine-grained control over each component should still evaluate Vapi.

§Latency Reality Check

Retell publishes a 250 to 400ms first-token-to-first-audio latency for the standard inbound configuration. This is competitive with the best Vapi configurations and reflects Retell's tight integration of Deepgram streaming STT with their orchestration layer.

The latency band widens for outbound calls because the carrier setup time (Twilio's outbound dial sequence) adds 1 to 3 seconds before the AI's first turn. For inbound qualification this is invisible; for outbound cold call openings, the prospect may hear silence at the start of the call and hang up before the AI speaks. Mitigations: pre-recorded greeting, ringback before answer detection, optimised SIP trunk routing.

Premium LLM choices (GPT-4o full, Claude Sonnet 3.5/3.6) add 200 to 500ms to the model-reasoning portion of latency. For complex objection handling this is worth the latency hit; for simple appointment-booking the budget LLM at sub-300ms total is the better experience trade-off.

§Retell vs Bland vs Vapi: Pricing Summary

PlatformHeadline rateAll-in typicalMonthly minimumStrength
Retell AI$0.07/min bundled$0.085-$0.20/minNoneBest bundle value; lowest latency
Vapi$0.05/min platform$0.25-$0.33/minNoneBest component flexibility
Bland AI$0.09-$0.14/min$0.10-$0.30/minPlan-basedBuilt-in dialer + sequencing
Synthflow$375-$900/mo + $0.08/min$0.08-$0.23/minPlan-basedNo-code friendly; flat plans

For pure cost-per-minute on sales workloads, Retell wins on most realistic configurations. The full head-to-head is at Vapi vs Retell.

§FAQ

Is the $0.07 per minute Retell rate real or marketing?
Real, with caveats. The $0.07 per minute is genuinely the bundled platform plus STT plus LLM plus TTS rate that Retell publishes and charges. It does not include telephony (an additional $0.014 to $0.03 per minute), premium LLM uplift if you choose GPT-4o full or Claude Sonnet, or premium TTS uplift if you choose ElevenLabs Multilingual v2. A bare-stack call at $0.07 plus Telnyx telephony at $0.005 lands at $0.075 per minute all-in, which is the lowest published all-in rate on any major voice AI platform in 2026.
Can I use my own OpenAI API key with Retell?
Bring-your-own-key on Retell is supported for some providers but is more constrained than Vapi. Default LLM selection is from Retell's curated provider list at bundled pricing. Custom BYOK is available on enterprise tier and requires direct sales contact. Most production deployments use the default LLM choices because the bundle saving outweighs the BYOK arbitrage.
What is Retell's free trial?
Retell offers a free $10 to $30 credit on new accounts for development and testing. The free credit is enough to build and test a working agent over a few hundred test calls. Production traffic moves to pay-as-you-go immediately after the credit is consumed.
Does Retell handle inbound and outbound differently in pricing?
The Retell bundled rate is the same for inbound and outbound. The telephony pass-through differs: outbound is $0.014 per minute on Twilio default; inbound is $0.0085 per minute plus a monthly per-number fee. For high-volume inbound, the lower per-minute telephony makes Retell even cheaper than the published $0.085 mid-tier figure.
Is Retell HIPAA compliant for healthcare?
Retell offers a HIPAA tier with BAA execution for healthcare workloads. The LLM, STT, and TTS providers in your stack must also be HIPAA-eligible (OpenAI, Anthropic, Google, Deepgram, ElevenLabs offer HIPAA enterprise tiers). The default pay-as-you-go pricing assumes non-PHI workloads; healthcare deployments require enterprise contract execution.
How does Retell handle voice cloning?
Retell supports custom voice cloning through ElevenLabs and PlayHT integrations. Voice cloning subscription costs apply separately (ElevenLabs Voice Lab at $99 to $1,320 per month; PlayHT Studio at $39 to $99 per month) on top of per-minute TTS. Most production sales deployments use a default ElevenLabs voice rather than a custom clone.

Updated 2026-05-11