Retell AI Pricing 2026: Plans, Costs & Alternatives

Retell AI charges a per-minute usage rate starting at $0.08 per minute on its Growth plan, with platform fees and telephony costs billed separately; this article breaks down the full cost structure, hidden fees, and how to estimate your monthly bill.

Major takeaways

How much does Retell AI cost? Retell AI's published pricing starts at $0.08 per minute on the Growth plan, with a $99/month platform fee; telephony, LLM inference, and premium voice models are billed separately based on usage.

What's included in the per-minute rate? The base per-minute rate covers Retell AI's platform orchestration, speech-to-text, and text-to-speech using standard voices; telephony, LLM costs, and premium voice models are separate line items on your invoice.

What else should you budget for? Telephony costs (typically $0.01–$0.02 per minute), phone number rentals ($1–$5/month per number), LLM inference fees (varies by model), and engineering time to build and maintain conversation flows add to the headline rate.

Quick answer: Retell AI pricing at a glance

What Retell AI is

Retell AI is an AI voice agent platform for developers and operations teams building conversational phone agents. The platform handles inbound and outbound calls, integrates with telephony providers, and supports custom conversation flows through a visual builder and API. Retell AI is used by contact centres, sales teams, and support operations to automate qualification, appointment setting, and customer service calls.

The platform connects to third-party LLM providers (OpenAI, Anthropic, Google) and telephony carriers (Twilio, Vonage, Bandwidth). Agents are configured through a dashboard or API, with conversation logic defined in JSON or through the visual flow editor. Retell AI handles speech-to-text, text-to-speech, and call orchestration; the buyer supplies the LLM and telephony accounts. The platform is sold on a usage basis with monthly platform fees and per-minute charges for call time.

How Retell AI pricing works

The pricing model

Retell AI uses a hybrid pricing model: a monthly platform fee plus per-minute usage charges. The platform fee grants access to the dashboard, API, and core features; the per-minute rate is billed for every minute of call time (inbound or outbound) that the agent handles. Telephony costs and LLM inference are billed separately through your own provider accounts, not through Retell AI.

As listed on Retell AI's pricing page at time of writing, the Growth plan charges $99/month plus $0.08 per minute; the Pro plan charges $299/month plus $0.06 per minute. Volume discounts are available on annual contracts. Enterprise pricing is custom and typically includes dedicated support, SLA guarantees, and negotiated per-minute rates for high-volume deployments. The platform fee is charged regardless of usage; if you run zero calls in a month, you still pay the base subscription.

What the per-minute rate covers

The per-minute rate covers Retell AI's platform orchestration, speech-to-text (STT), and text-to-speech (TTS) using standard voices. STT is powered by Deepgram or AssemblyAI (configurable); TTS uses ElevenLabs, PlayHT, or Cartesia voices at the standard tier.

The rate does not include telephony (you pay your carrier directly), LLM inference (you pay OpenAI, Anthropic, or Google directly), or premium voice models (billed as an add-on).

Retell AI's documentation specifies that the per-minute charge starts when the call connects and ends when the call terminates. Unanswered calls, busy signals, and voicemail drops are not billed by Retell AI, but your telephony provider may still charge for the attempt. The platform does not charge separately for concurrent calls; concurrency limits are set by plan tier, not priced per seat.

Limits and add-ons

The Growth plan includes up to 10 concurrent calls; the Pro plan includes up to 50 concurrent calls. Additional concurrency is available as an add-on at $50/month per 10 concurrent calls, as listed on the pricing page. Phone number rentals are billed separately through your telephony provider (Twilio charges $1/month per number; Vonage charges $1–$5/month depending on country).

Premium voice models (ultra-realistic voices from ElevenLabs or PlayHT) add $0.02–$0.04 per minute to the base rate, depending on the voice selected. LLM costs vary by model: OpenAI's GPT-4 Turbo costs approximately $0.01–$0.03 per call minute depending on conversation complexity; Anthropic's Claude 3.5 Sonnet costs $0.015–$0.025 per call minute. These are external costs billed by the LLM provider, not Retell AI. The platform does not bundle LLM credits.

Plan breakdown

Growth plan. The entry tier for teams testing voice agents or running low-volume campaigns. The $99/month platform fee is fixed; the $0.08/min rate applies to every call minute. At 1,000 minutes/month, total Retell AI charges are $179/month (platform fee + usage). At 5,000 minutes/month, total charges are $499/month.

The 10-concurrent-call limit means you can handle 10 simultaneous conversations; if an 11th call arrives, it queues or fails depending on your telephony configuration. Standard voices are included; premium voices add $0.02–$0.04/min. Telephony and LLM costs are external.

Pro plan. Designed for teams running consistent call volume. The $299/month platform fee buys a lower per-minute rate ($0.06 vs $0.08) and 50 concurrent calls. At 10,000 minutes/month, total Retell AI charges are $899/month (platform fee + usage). At 25,000 minutes/month, total charges are $1,799/month. Priority support includes faster response times and a dedicated Slack channel. Custom integrations (CRM webhooks, calendar APIs) are available on this tier. The break-even point where Pro becomes cheaper than Growth is around 10,000 minutes/month.

Enterprise plan. Custom pricing negotiated based on volume commitments, use case, and compliance requirements. Retell AI's sales team typically asks for projected monthly minutes, peak concurrency needs, and whether you need SOC 2 compliance or data residency guarantees. Enterprise contracts often include volume discounts (per-minute rates as low as $0.04/min for commitments above 100,000 minutes/month, based on third-party reports), dedicated customer success, and SLA-backed uptime guarantees.

Minimum contract length is typically 12 months.

Estimate your monthly cost

The formula for estimating your monthly Retell AI bill:

Monthly cost ≈ (calls per month × average minutes per call × all-in per-minute rate) + platform fee + phone number fees + add-ons

Each input:

• Calls per month: Total inbound and outbound calls your agents handle. For outbound, multiply dials by connect rate (typically 5–15% for cold outreach, 40–60% for warm follow-ups). For inbound, use your current call volume or projected traffic.
• Average minutes per call: Measured from connect to hangup. Qualification calls average 2–4 minutes; appointment-setting calls average 3–6 minutes; support calls average 5–10 minutes. Include hold time and transfers if your flow supports them.
• All-in per-minute rate: Retell AI's base rate ($0.06–$0.08/min) + telephony ($0.01–$0.02/min) + LLM inference ($0.01–$0.03/min) + premium voices if used ($0.02–$0.04/min). Total typically $0.10–$0.17/min.
• Platform fee: $99/month (Growth) or $299/month (Pro).
• Phone number fees: $1–$5/month per number. Most teams rent 1–10 numbers.
• Add-ons: Extra concurrency ($50/month per 10 calls), premium integrations, dedicated support.

Illustrative scenarios

The table below uses illustrative assumptions for call volume, length, and rates. Replace these with your own call data to estimate your actual cost.

Telephony estimate assumes $0.02/min; LLM estimate assumes $0.03/min (GPT-4 Turbo). Your actual costs depend on carrier rates, model choice, and conversation complexity. If you use premium voices, add $0.02–$0.04/min to the Retell AI usage cost.

Hidden and adjacent costs

Telephony and phone number fees

Retell AI does not include telephony in its pricing. You pay your carrier (Twilio, Vonage, Bandwidth) directly for inbound and outbound minutes. Twilio charges $0.0085/min for US inbound calls and $0.013/min for US outbound calls; international rates range from $0.02/min (Canada, UK) to $0.30/min (some African and Asian countries). Phone number rentals cost $1/month per US number on Twilio; toll-free numbers cost $2/month.

If you run 10,000 minutes/month on Twilio, budget $100–$200/month for telephony alone.

LLM and voice model upgrades

LLM inference is billed by your LLM provider, not Retell AI. OpenAI's GPT-4 Turbo costs $0.01 per 1,000 input tokens and $0.03 per 1,000 output tokens; a 3-minute conversation typically consumes 2,000–4,000 tokens total, costing $0.02–$0.04 per call. Anthropic's Claude 3.5 Sonnet costs $0.003 per 1,000 input tokens and $0.015 per 1,000 output tokens, averaging $0.015–$0.025 per call. Premium voice models (ElevenLabs ultra-realistic voices) add $0.02–$0.04/min to your Retell AI bill.

Engineering time to build and maintain flows

Building a production-ready voice agent requires engineering time to design conversation flows, handle edge cases, and integrate with your CRM or calendar. Initial setup typically takes 20–40 hours for a simple qualification agent; complex multi-step flows (appointment setting with calendar integration, escalation to humans) can take 80–120 hours. Ongoing maintenance (updating scripts, fixing bugs, tuning LLM prompts) averages 5–10 hours/month.

At a $100/hour engineering rate, budget $2,000–$4,000 for initial build and $500–$1,000/month for maintenance.

Integration costs

Retell AI integrates with CRMs (Salesforce, HubSpot), calendars (Google Calendar, Calendly), and helpdesks (Zendesk, Intercom) via webhooks and API calls. Most integrations require custom code to map data fields and handle error states. If your CRM doesn't have a pre-built connector, expect 10–20 hours of engineering time to build the integration. Some CRMs charge API usage fees; Salesforce charges $0.002 per API call above the included limit, which can add $50–$200/month if your agents log every call.

QA and call review

Voice agents require ongoing QA to catch errors, tune responses, and improve conversion rates. Most teams review 5–10% of calls manually; at 10,000 calls/month, that's 500–1,000 calls to review. Listening to a 3-minute call and logging feedback takes 5–7 minutes; reviewing 500 calls takes 40–60 hours/month. At a $25/hour QA rate, budget $1,000–$1,500/month for call review.

Retell AI's dashboard includes call recordings and transcripts; you supply the labour to review them.

Compliance and data retention

Voice agents that handle personal data (names, phone numbers, health information, payment details) must comply with GDPR, CCPA, HIPAA, or PCI-DSS depending on jurisdiction and use case. Retell AI is SOC 2 Type II compliant, but compliance work (consent capture scripts, data retention policies, audit logs) falls on your team. Legal review of call scripts and privacy policies typically costs $2,000–$5,000 upfront.

Data retention (storing call recordings for 1–7 years) adds storage costs; at $0.02/GB/month, 10,000 calls (3 minutes each, ~5MB per recording) cost $10/month to store.

Is Retell AI worth the price?

Retell AI delivers value for teams that need a developer-friendly platform with flexible LLM and telephony integrations. The per-minute rate is competitive with Bland AI and Vapi when you factor in the included STT/TTS; the platform fee is lower than Vapi's $500/month enterprise minimum. Teams running 5,000+ minutes/month on the Pro plan pay an all-in cost (Retell AI + telephony + LLM) of $0.12–$0.17/min, which is 60–75% cheaper than hiring human SDRs or receptionists at $20–$30/hour.

The platform is less cost-effective for very low-volume use cases (under 1,000 minutes/month). At 500 minutes/month, the $99 platform fee plus $40 in usage charges totals $139/month, or $0.28/min effective rate once you add telephony and LLM costs. Teams running sporadic campaigns (a few hundred calls per quarter) pay more per call than they would with a pay-as-you-go platform like Bland AI, which has no platform fee.

Public review volume for Retell AI is limited; the recurring themes below are drawn from the available G2 / Trustpilot / TrustRadius coverage. Users on G2 note that the platform requires engineering resources to set up and maintain; non-technical teams report needing to hire a developer or agency to build flows. The API-first design is praised by technical buyers but cited as a barrier by operations teams without in-house engineering.

Some reviewers flag that LLM costs can spike unexpectedly if conversation flows aren't optimised; one user reported a $400 LLM bill in a single month due to overly verbose agent responses.

Retell AI alternatives to price against

Bland AI charges $0.09/min with no platform fee, making it cheaper than Retell AI for low-volume use cases (under 2,000 minutes/month). Bland AI bundles telephony and LLM costs into the per-minute rate, simplifying billing. The platform is less flexible than Retell AI for custom integrations; it's best for teams running simple qualification or appointment-setting flows. Bland AI's pricing is detailed in Bland AI Pricing 2026: Per-Minute Costs & Plan Breakdown.

Vapi charges $0.05/min on its Enterprise plan, which requires a $500/month minimum. The break-even point where Vapi becomes cheaper than Retell AI Pro is around 15,000 minutes/month. Vapi supports white-label deployments and custom voice models, making it a fit for agencies and high-volume contact centres. The platform requires more engineering work than Retell AI to set up. A detailed comparison is available in Retell AI vs Vapi: Honest Comparison (2026).

Synthflow charges $0.12/min all-in, including telephony, STT, TTS, and LLM inference. The bundled pricing simplifies budgeting but offers less control over LLM and telephony providers. Synthflow is best for teams that want a single invoice and don't need custom integrations. The platform is newer and has limited third-party review coverage.

11x (Julian) is an AI phone agent for inbound sales, sold as part of the 11x platform alongside Alice (AI SDR for outbound). Julian handles inbound speed-to-lead, qualification, and routing to human reps. Pricing is custom and based on call volume and platform scope; 11x does not publish per-minute rates. Julian is a fit for revenue teams that want inbound voice and outbound email/LinkedIn in a single system. It's not a standalone voice platform like Retell AI.

Air.ai offers custom enterprise pricing for contact centres running 24/7 support or sales operations. The platform is designed for high-volume deployments (50,000+ minutes/month) and includes dedicated account management. Air.ai does not publish pricing; third-party estimates suggest per-minute rates of $0.08–$0.15/min on annual contracts.

How to negotiate with Retell AI

Commit to volume for a lower per-minute rate

Retell AI offers volume discounts on annual contracts. Committing to 50,000+ minutes/month can reduce the per-minute rate from $0.06 to $0.04–$0.05/min, based on third-party reports. Ask the sales team for a tiered pricing schedule that lowers the rate as you scale. Lock in the discount for the contract term to avoid rate increases.

Ask what's included in the per-minute rate

Clarify whether the quoted rate includes STT, TTS, telephony, and LLM costs, or if those are billed separately. Retell AI's published rates cover STT and TTS but exclude telephony and LLM; confirm this in writing to avoid surprise invoices. If you're comparing vendors, normalise rates to an all-in cost (platform + telephony + LLM) for an accurate comparison.

Cap overage charges

If you're on a volume-committed plan, negotiate a cap on overage charges. Some vendors charge 2–3× the base rate for minutes above the committed volume; Retell AI's standard overage rate is not publicly listed. Ask for overage to be billed at the same per-minute rate as committed volume, or negotiate a grace buffer (e.g., 10% overage at no penalty).

Negotiate concurrency limits

The Pro plan includes 50 concurrent calls; additional concurrency costs $50/month per 10 calls. If your peak concurrency is 60–70 calls, negotiate a higher included limit (e.g., 75 concurrent calls) instead of paying for add-ons. Concurrency limits are often flexible on annual contracts.

Pilot before signing an annual contract

Run a 1–3 month pilot on the Growth or Pro plan before committing to an annual contract. Use the pilot to validate call quality, measure LLM costs, and test integrations. If the pilot uncovers issues (high LLM costs, poor STT accuracy, integration gaps), you can switch vendors without breaking a long-term contract.

Confirm data retention and compliance terms

Ask how long Retell AI retains call recordings and transcripts, and whether you can configure retention periods. Some industries (healthcare, finance) require 7-year retention; others (marketing) prefer 30–90 days to minimise storage costs. Confirm whether Retell AI is HIPAA-compliant or PCI-DSS-compliant if you handle sensitive data; these certifications may require an enterprise plan.

Verdict

Retell AI's pricing suits technical teams running 5,000–50,000 minutes/month that need flexible LLM and telephony integrations. The Pro plan's $0.06/min rate plus $299/month platform fee delivers an all-in cost of $0.12–$0.17/min when you add telephony and LLM, which is competitive for mid-market deployments. The platform's API-first design and support for multiple LLM providers make it a fit for teams with engineering resources.

Teams running under 2,000 minutes/month should compare Bland AI, which has no platform fee and bundles telephony into a $0.09/min rate. Non-technical teams without in-house developers should evaluate platforms with visual builders and managed services, as Retell AI requires code to build and maintain flows. High-volume contact centres (50,000+ minutes/month) should request enterprise quotes from Retell AI, Vapi, and Air.ai to compare volume discounts and SLA terms.

FAQ

How much does Retell AI cost per minute?

Retell AI charges $0.08/min on the Growth plan and $0.06/min on the Pro plan, as listed on its pricing page. These rates cover platform orchestration, STT, and TTS with standard voices. Telephony, LLM inference, and premium voices are billed separately. All-in cost (Retell AI + telephony + LLM) typically ranges from $0.10/min to $0.17/min depending on your carrier and LLM provider.

Does Retell AI have a free trial?

Retell AI offers $50 in free credits on signup, which covers approximately 600 minutes of call time at the $0.08/min Growth rate (excluding telephony and LLM costs). The credits expire after 30 days. No credit card is required to start the trial. After the credits are exhausted, you must add a payment method to continue using the platform.

Is telephony included in Retell AI's pricing?

No. Retell AI's per-minute rate covers STT, TTS, and platform orchestration; telephony is billed separately through your carrier (Twilio, Vonage, Bandwidth). You connect your own telephony account to Retell AI via API. Twilio charges $0.0085–$0.013/min for US calls; international rates vary. Phone number rentals ($1–$5/month per number) are also billed by your carrier, not Retell AI.

Does Retell AI charge for failed or unanswered calls?

Retell AI does not charge for unanswered calls, busy signals, or voicemail drops; the per-minute rate applies only to connected call time. However, your telephony provider may still charge for the dial attempt. Twilio charges $0.013/min for outbound calls regardless of whether the call connects; confirm your carrier's billing policy to understand the full cost of failed calls.

Is there a minimum contract?

The Growth and Pro plans are billed monthly with no minimum contract; you can cancel at any time. Enterprise plans typically require a 12-month contract with a committed monthly minute volume. Annual contracts on Growth or Pro may include discounts (5–10% off the monthly rate, based on third-party reports) but lock you in for the full term.

How does Retell AI pricing compare to hiring a human receptionist or SDR?

A human receptionist or SDR costs $20–$30/hour (salary + benefits + overhead), or $3,200–$4,800/month for full-time work. At 160 hours/month, that's $20–$30 per productive hour. Retell AI's all-in cost (platform + telephony + LLM) is $0.12–$0.17/min, or $7.20–$10.20/hour of call time. If your agent handles 30 calls/hour at 2 minutes each, the cost per hour is $7.20–$10.20 vs $20–$30 for a human. The break-even point depends on call volume and complexity; simple qualification calls favour AI, while complex sales conversations may still require human reps. For a detailed cost comparison, see the SDR cost calculator.

What's the cheapest alternative to Retell AI?

Bland AI is the cheapest alternative for low-volume use cases, charging $0.09/min with no platform fee. At 1,000 minutes/month, Bland AI costs $90/month vs Retell AI's $179/month (Growth plan). Bland AI bundles telephony and LLM into the per-minute rate, simplifying billing. The platform is less flexible than Retell AI for custom integrations and LLM choice. For details, see Bland AI Pricing 2026: Per-Minute Costs & Plan Breakdown.

How does Retell AI compare to 11x's Julian?

Julian is an AI phone agent for inbound sales, sold as part of the 11x platform alongside Alice (AI SDR for outbound email and LinkedIn). Julian handles inbound speed-to-lead, qualification, and routing to human reps; pricing is custom and based on call volume and platform scope. Retell AI is a standalone voice platform with published per-minute pricing and support for custom LLM and telephony integrations. Julian suits revenue teams that want inbound voice and outbound prospecting in a single system; Retell AI suits technical teams building custom voice agents across multiple use cases.

Can I use Retell AI without engineering resources?

Retell AI's API-first design requires code to build and maintain conversation flows. Non-technical teams typically hire a developer or agency to set up the platform. If you need a no-code solution, evaluate platforms with visual builders and managed services. Retell AI is best for teams with in-house engineering or budget to outsource development.

What happens if I exceed my concurrency limit?

If you exceed your plan's concurrency limit (10 concurrent calls on Growth, 50 on Pro), additional calls queue or fail depending on your telephony configuration. You can purchase additional concurrency at $50/month per 10 concurrent calls. On Enterprise plans, concurrency limits are negotiated based on your peak traffic needs.

Last updated: October 2026.