Quick answer: As at 31 October 2026, Retell AI (retellai.com) charges pay-as-you-go voice agents per minute, tracked to the second and paid from prepaid credits: $0.055 for its voice engine, plus your LLM ($0.0016 to $0.64), your voice ($0.015 to $0.10) and telephony ($0.015 on Retell’s US Twilio numbers). Our three worked stacks cost $95 to $342 for 1,000 minutes and $1,870 to $6,810 for 20,000. Twenty concurrent calls are included.
This page is about Retell AI, the voice and chat agent platform at retellai.com. Every figure comes from Retell’s own pricing page, documentation, changelog and terms of service, read on 31 October 2026. Retell prints a bare dollar sign and its documentation reports costs in USD, so every figure is in US dollars and excludes tax, including Australian GST. Enterprise contracts are quoted privately and are not priced here.
How much does Retell AI cost?
Retell sells two plans. Pay-as-you-go has no plan fee: “Retell starts at $0, pay only for what you use”, and the page promises “No minimums, no contracts, no surprises.” The Enterprise Plan is “Custom Pricing”, “with volume pricing that scales as you grow”, dedicated support and a higher cap on concurrent calls. The pay-as-you-go headline is “$0.07-$0.31 / min for AI Voice Agents” and “$0.002+ / msg for AI Chat Agents”.
That headline range is a summary, not a price quote. A Retell voice minute is the sum of four lines: Retell’s voice engine (Voice Infra), a language model from OpenAI, Anthropic or Google, a voice from Retell, ElevenLabs, Cartesia, Minimax, Fish, OpenAI or Inworld, and telephony over Retell’s Twilio or Telnyx numbers or your own SIP trunk. Add-ons and monthly items sit on top.
| Price element | What Retell’s own page says, as at 31 October 2026 | Where it says it |
|---|---|---|
| Plans | Pay as you go (from $0, no plan fee) and Enterprise (custom pricing) | Pricing page |
| Platform fee per minute | Retell Voice Infra $0.055/minute | Pricing page, Conversation Voice Engine |
| Language model per minute | 35 models plus an Ultrafast option, from $0.0016 (GPT 5 nano) to $0.32 (GPT 6 Astra) on the standard tier; Fast Tier up to $0.64 | Pricing page, LLM table |
| Voice (text-to-speech) per minute | $0.015 for Retell Platform, Minimax, Fish, Cartesia, OpenAI and Inworld voices; ElevenLabs $0.040 (Flash) to $0.10 (v3, v4) | Pricing page, Text-to-speech table |
| Telephony per minute | US $0.015 on Retell’s Twilio numbers; no charge for SIP trunking or custom telephony; other countries $0.03 to $0.80 | Pricing page; International calling doc |
| Phone numbers | $2.00 a month (US, Canada); US toll-free $5 a month | Pricing page; Purchase phone number doc |
| Concurrency | 20 concurrent calls free; $8.00 per extra concurrent call per month | Pricing page FAQ and Monthly Subscription Items |
| Free credit | $10 in free credits for new accounts, once per email address | Pricing page; Billing doc |
| Billing basis | Prepaid credits deducted in real time; monthly items billed to your card at month end | Billing doc |
| Refunds | Credits “non-refundable once purchased”; fees non-refundable except as required by law | Billing doc; Terms of Service 2.1 |
| Enterprise | Custom pricing; custom concurrency “starting at 50+ calls”; stays on monthly invoicing | Pricing page; Billing doc |
So the honest answer to “how much is Retell AI per month” is: nothing until you make calls, then $0.093 to $0.340 a minute on the three stacks we work through below, plus a few dollars a month per phone number. The rest of this page shows where in that span your bill lands.
Where Zian fits
If you are pricing Retell because you want AI agents that answer and make your calls, put Zian beside it: Zian is an AI sales and reception agents platform that answers inbound calls 24/7 as an AI receptionist and runs outbound follow-up by voice, SMS, email, WhatsApp, Instagram DMs and Facebook Messenger. Retell is a platform your team builds agents on and pays for by the minute; Zian’s agents qualify, book, transfer and take messages as a ready digital team, and Zian prints no price today because it is in partnership-application beta, where each of its three listed plans, including Intel + Widgets (Free Forever), starts with a partnership application.
Zian’s selling layer is published by name. SmartReach AI™ sets the message, channel, timing and follow-up pacing, and PrecisionPitch AI™ keeps split-testing scripts against real outcomes. Its agents handle answering machines and voicemail on calls, work in 30+ languages with voice cloning, run research, web and knowledge-base lookups, connect to HubSpot, Salesforce, HighLevel and Zapier, and can be deployed as a private model on customer infrastructure. Zian’s live counter shows 50,769+ qualified sales appointments set.
If you would rather not run any platform yourself, LeadsNow.ai is the done-for-you option: a managed AI appointment setting and lead generation service that runs on Zian, priced on a pay-per-result basis, with leads answered in under 60 seconds, 24/7.
Want the receptionist and the outbound follow-up on one platform? Zian is in partnership-application beta. Apply For Partnership
What you are billed for on Retell AI
Retell’s bill has three kinds of line. Usage is “deducted from your credit balance in real time”; phone numbers, extra knowledge bases, calls per second (CPS) upgrades, purchased concurrency and SMS subscriptions are “billed to your payment method on file at the end of each billing cycle”; and a few fees are one-off. This is every line on the pricing page plus the fees we found in Retell’s documentation, as at 31 October 2026.
| Line | Price | How it is billed | Condition Retell states |
|---|---|---|---|
| Retell Voice Infra | $0.055/min | Credits, per minute | Every voice call; Retell calls it “the flat per-minute platform cost of running the real-time call” |
| LLM | $0.0016 to $0.64/min | Credits, per minute | Price depends on the model and on Standard, Fast or Ultrafast tier |
| Text-to-speech | $0.015 to $0.10/min | Credits, per minute | ElevenLabs models are $0.040 to $0.10; six other voice families are $0.015 |
| Telephony | $0.015/min (US) to $0.80/min (Philippines) | Credits, per minute | No charge for SIP trunking or custom telephony; toll-free inbound $0.06/min; Retell’s Telnyx US route is $0.03/min |
| Knowledge Base | +$0.005/min | Credits, per minute | On calls with a knowledge base enabled, however many it uses |
| Advanced Denoising | +$0.005/min | Credits, per minute | Docs tie it to the “Remove noise + background speech” mode |
| Safety Guardrails | +$0.005/min | Credits, per minute | Optional add-on |
| PII Removal | +$0.01/min | Credits, per minute | Optional add-on |
| AI Quality Assurance | $0.10/min of analysed call time | Credits | First 100 minutes free per workspace; sampling controls volume |
| Concurrency burst | +$0.10/min | Credits | Applies to the entire call if it started above your limit |
| Batch Call | +$0.005/dial | Credits, per dial | Batch calls are now deprecated in favour of the Dialer but still available |
| Branded Call | +$0.10/outbound call | Per outbound call | US numbers only; on stopping, billing runs to the end of the cycle |
| Conversation Memory | $0.005 per run | Per save; a backfill is $0.005 per conversation processed | Announced 8 October 2026 |
| Conductor (build assistant) | $0.20 per credit | Billed to the workspace, per credit | Free up to 30 credits per user and 200 per workspace a day; billed only if an admin turns on pay-as-you-go |
| SMS messages | $0.01/msg | Credits, per message | US numbers only, not toll-free, not Telnyx numbers |
| Chat agent messages | $0.0005 to $0.10 per AI message | Credits, per message | Price depends on the model |
| Retell phone number | $2.00/month; toll-free $5/month | Card, monthly, prorated | No charge for custom phone numbers; bills until released |
| Retell SMS | $20.00/month per number | Card, monthly | No charge for custom telephony SMS |
| Extra concurrency | $8.00 per concurrent call per month | Card, monthly | First 20 free; charged immediately when you raise the limit |
| Extra knowledge bases | $8.00 per knowledge base per month, prorated | Card, monthly | First 10 free (“$80/month free value”) |
| Verified Phone Number | $10.00 per number per month | Monthly | US numbers only; one tab of the pricing page adds “One-time $10 fee” |
| Daily spam check | $2 per number per month | Monthly | US numbers only (Dialer docs, as of October 2026) |
| SMS registration (A2P 10DLC) | Brand $4 (low-volume) or $45 (standard); campaign $15 | One-off | Campaign fee non-refundable even if rejected |
Two cells are less tidy than the summary. The verified-number row reads “$10.00/Phone number/month” on the pricing page’s default tab, while the same row on the Speech-to-Speech tab adds the note “One-time $10 fee”; Retell’s Dialer documentation says “Verified numbers cost $10 per month each”, so budget it monthly and ask Retell whether a one-time fee applies as well before you budget a large pool. And the Batch Call line is still on the pricing page, but Retell’s documentation says “Batch calls are deprecated” and points new campaigns to the Dialer. We could not find a per-dial fee for Dialer campaigns on the pricing page, the Dialer overview, the caller-ID and pacing page or the compliance-checks page, as at 31 October 2026. Do-not-call screening in the Dialer is “billed per number checked”, priced only in the dashboard estimate.
Which model and voice you pick sets most of the Retell bill
Voice Infra ($0.055) and US telephony on Retell’s Twilio numbers ($0.015) fix $0.07 of every US minute; the rest is the model and the voice. Retell’s LLM table lists 35 models plus an Ultrafast option, each priced per minute rather than per token, with a second column for Fast Tier. Retell’s LLM documentation describes Fast Tier as “a higher-priority resource pool” and says the price increase “depends on the model and tier; Fast does not use one fixed multiplier across all models.” A selection, quoted from the table:
| Model (Retell’s label) | Standard, per minute | Fast Tier, per minute |
|---|---|---|
| GPT 5 nano (cheapest listed) | $0.0016 | not listed |
| GPT 5 mini | $0.008 | $0.016 |
| Gemini 3.0 Flash | $0.016 | not listed |
| Claude 4.5 haiku | $0.025 | not listed |
| GPT 5.1 | $0.04 | $0.08 |
| GPT 5.6 Terra (marked Recommended) | $0.064 | not listed |
| Claude 5 Sonnet (marked Recommended) | $0.064 | not listed |
| GPT 5.4 (marked Recommended) | $0.080 | $0.16 |
| GPT 5.5 (marked Recommended) | $0.16 | $0.32 |
| Claude 5 Opus | $0.16 | not listed |
| GPT 6 Astra (dearest on the standard tier) | $0.32 | $0.64 |
| GPT 6.1 Sol (Ultrafast) | $0.384, a separate Ultrafast row | |
The spread is the biggest driver of a Retell bill: the dearest standard model costs 200 times the cheapest. The four “Recommended” models run from $0.064 to $0.16. Speech-to-speech models (GPT Realtime, $0.07 to $0.38 a minute) sit on a separate tab; the pages we read do not say how they combine with the other lines, so we have not modelled them.
Voices are simpler. Six families cost $0.015 a minute. ElevenLabs is priced by model: Flash $0.040, Multilingual v2 $0.06, v3 Conversational and v4 Turbo $0.08, and v3 and v4 $0.10. Moving from a $0.015 voice to ElevenLabs v3 adds $0.085 to every minute, which is more than the Voice Infra fee itself.
One model cost is easy to miss: Retell’s “fast fallback” runs a backup model alongside the primary and “can incur charges for both models, including when only one response is used.” Per-minute model pricing also stops shielding you from prompt growth above 4,000 tokens (see below). For why a long call’s model cost climbs on token-billed platforms, our breakdown of voice AI cost per minute for realtime and pipeline stacks walks through it.
Retell AI cost worked at 1,000, 5,000 and 20,000 minutes
This is the worked calculation. Every line is a published rate multiplied by a quantity you control, so you can substitute your own numbers.
Inputs, stated in full:
- Connected, billed minutes a month of 1,000, 5,000 and 20,000, all to and from US numbers, on Retell’s Twilio telephony at $0.015 a minute.
- Stack A (lean): GPT 5 mini ($0.008) and a Retell Platform voice ($0.015).
- Stack B (mid): GPT 5.6 Terra ($0.064, marked Recommended), ElevenLabs Flash ($0.040) and Knowledge Base on every call ($0.005).
- Stack C (premium): GPT 5.5 ($0.16, marked Recommended), ElevenLabs v3 ($0.10), Knowledge Base ($0.005) and Advanced Denoising ($0.005).
- Retell Voice Infra $0.055 on every stack, standard LLM tier, prompts under 4,000 tokens.
- Retell numbers at $2 a month: one at 1,000 minutes, two at 5,000, five at 20,000. One knowledge base, inside the free ten.
- Peak concurrency at or under 20. Spread over 22 weekdays of 8 hours, 20,000 minutes is about 1.9 calls in progress on average.
- No Enterprise plan, no Fast Tier, no SMS, no AI QA, no branded or verified numbers. The one-off $10 trial credit is ignored.
Per-minute rate for each stack: A is $0.055 + $0.008 + $0.015 + $0.015 = $0.093. B is $0.055 + $0.064 + $0.040 + $0.015 + $0.005 = $0.179. C is $0.055 + $0.16 + $0.10 + $0.015 + $0.010 = $0.340.
| Line item | A, 1,000 min | A, 5,000 min | A, 20,000 min | B, 1,000 min | B, 5,000 min | B, 20,000 min | C, 1,000 min | C, 5,000 min | C, 20,000 min |
|---|---|---|---|---|---|---|---|---|---|
| Voice Infra ($0.055) | $55.00 | $275.00 | $1,100.00 | $55.00 | $275.00 | $1,100.00 | $55.00 | $275.00 | $1,100.00 |
| LLM | $8.00 | $40.00 | $160.00 | $64.00 | $320.00 | $1,280.00 | $160.00 | $800.00 | $3,200.00 |
| Voice | $15.00 | $75.00 | $300.00 | $40.00 | $200.00 | $800.00 | $100.00 | $500.00 | $2,000.00 |
| US telephony ($0.015) | $15.00 | $75.00 | $300.00 | $15.00 | $75.00 | $300.00 | $15.00 | $75.00 | $300.00 |
| Add-ons | $0.00 | $0.00 | $0.00 | $5.00 | $25.00 | $100.00 | $10.00 | $50.00 | $200.00 |
| Phone numbers ($2 each) | $2.00 | $4.00 | $10.00 | $2.00 | $4.00 | $10.00 | $2.00 | $4.00 | $10.00 |
| Concurrency (20 or fewer) | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 | $0.00 |
| Monthly total | $95.00 | $469.00 | $1,870.00 | $181.00 | $899.00 | $3,590.00 | $342.00 | $1,704.00 | $6,810.00 |
| Effective cost per minute | $0.0950 | $0.0938 | $0.0935 | $0.1810 | $0.1798 | $0.1795 | $0.3420 | $0.3408 | $0.3405 |
How we recomputed it. Line by line, and again as rate × minutes + numbers: Stack A at 5,000 minutes is $0.093 × 5,000 + $4 = $469.00; Stack C at 20,000 is $0.340 × 20,000 + $10 = $6,810.00. Both methods agree on all nine columns.
What the worked bill says. On pay-as-you-go, volume does not lower Retell’s per-minute rate; only the number fees spread thinner. The stack moves the bill. At 20,000 minutes the premium stack costs $4,940 more a month than the lean one, and $4,740 of that gap is the model and voice lines. Stack C’s $0.340 a minute also sits above Retell’s own “$0.07-$0.31” headline, which is simply arithmetic: a Recommended model plus a top ElevenLabs voice plus two add-ons passes $0.31.
The same stacks change quickly outside the US. Calling Australian numbers over Retell’s Twilio route costs $0.10 a minute rather than $0.015, so Stack A becomes $0.178 a minute and 5,000 minutes costs $894 with two numbers ($0.178 × 5,000 + $4). Retell’s purchase-number page says it currently sells only US and Canadian numbers (its international-calling page mentions international numbers, so check the dashboard); for an Australian caller ID you would bring your own carrier, which removes Retell’s telephony line but adds your carrier’s bill. For a side-by-side on a different stack and volume, our Vapi vs Retell AI pricing comparison at 10,000 minutes prices Retell against a pass-through rival.
What outbound calling adds to a Retell bill
Assume 2-minute connected calls and 2.5 dials per connected call, so 1,000 minutes is 500 connected calls and 1,250 dials.
| Outbound line | 1,000 min | 5,000 min | 20,000 min | Basis |
|---|---|---|---|---|
| Batch Call at $0.005 per dial | $6.25 | $31.25 | $125.00 | 1,250 / 6,250 / 25,000 dials |
| Branded Call at $0.10, connected calls only | $50.00 | $250.00 | $1,000.00 | 500 / 2,500 / 10,000 calls |
| Branded Call at $0.10, every dial | $125.00 | $625.00 | $2,500.00 | 1,250 / 6,250 / 25,000 dials |
| Stack A with Batch Call, total | $101.25 | $500.25 | $1,995.00 | Worked total above + Batch line |
Retell prices Branded Call “+$0.10/outbound call” and does not say on the pages we read whether an unanswered dial counts, so the table shows both ends. At the every-dial end, branded caller ID on a lean stack costs more than the model and voice combined.
Throughput is set by concurrency: Dialer campaign calls “never use burst; they wait for a free slot instead”, and extra slots are $8 a month each, so 30 concurrent calls cost $80 a month above the free 20. If you are sizing the campaign layer itself, our Synthflow vs Retell AI outbound calling comparison covers the batch and concurrency mechanics, as checked in September 2026. Adding texting is $20 a month per number plus $0.01 a message, after one-off A2P 10DLC registration fees of $4 or $45 for the brand and $15 for the campaign; our guide to A2P 10DLC costs for two-way AI SMS agents explains the carrier side of those fees.
Running outbound and weighing build against buy? Zian’s agents run the calls, SMS, email and WhatsApp follow-up as one sequence, with SmartReach AI™ pacing it. Apply For Partnership
How Retell bills you: prepaid credits since 30 September 2026
Retell pay-as-you-go is now prepaid. Its billing documentation says new accounts pay “for usage upfront with prepaid credits”, and that workspaces still on the old monthly invoices were moved to credits on 30 September 2026, with the balance starting at $0 for anyone who had not switched. Enterprise workspaces stay on monthly invoicing. The rules that matter for a budget:
- Usage stops at zero. “When your credit balance reaches zero, calls, chat, and other usage stop until you buy credits or auto recharge tops up your balance.” Calls cut this way show a credit_exhausted reason.
- Auto recharge has bounds. As of October 2026 the trigger can be $10 to $1,000 and the restored balance $20 to $5,000, and the restored balance must be higher than the trigger.
- An optional monthly budget caps spend. Where the workspace shows Budget Setting, you can set $10 to $100,000; at 100% “calls and API requests pause”, even with credit left.
- Monthly items still hit your card. Numbers, extra knowledge bases, CPS upgrades and purchased concurrency are invoiced at the end of each calendar month and prorated if bought mid-month.
Refunds are the limb most cost pages leave out. Retell’s billing page says “Credits never expire, but they are non-refundable once purchased”, and its terms of service (last updated 25 October 2026) say “All fees are non-refundable except as required by law or as otherwise expressly provided for in these Terms.” The terms add three points. To cancel, delete the workspace or remove all subscription items, such as phone numbers, before the period ends, or the services renew and you owe the renewal fees. A failed automatic payment becomes an invoice due in 14 days plus a 7-day grace period, with a late fee of the lesser of 10% of the past-due amount or the legal maximum, and Retell may suspend or stop the services if payment has not arrived by the end of the grace period. And Retell “may change the rates at any time” by posting them on its pricing page, effective from your next billing cycle.
Against all this stands Retell’s own promise, “No minimums, no contracts, no surprises”, which is accurate: the plan has no fee and no term. The prepaid balance is the nearest thing to a commitment, so size top-ups to a few weeks of usage.
The billing rules that change your minutes
Retell publishes rules that change your billed minutes; its calculator does not show them.
- Seconds, not rounded minutes. “Each call is tracked to the nearest second … No rounding up per call.”
- Silence is billed. “Billing covers the entire duration of the call because the speech-to-text engine remains active and listening throughout, even during silence.”
- Failed calls are free; voicemail is partly billed. “You are only charged for connected calls … For calls that reach voicemail, billing applies only for the duration the AI agent is active on the line.”
- Transfers drop the agent fee. After a transfer, “Only the telephony fee continues for the remainder of the transferred call.”
- A 10-second minimum. Calls under 10 seconds that use a dynamic opening message when the AI speaks first are billed as 10 seconds.
- Long prompts scale the bill. Above 4,000 tokens, “Billed Duration = Original Duration × Scaling Factor (rounded up)”, where the factor is prompt tokens ÷ 4,000. Retell’s example is a 60-second call with 4,800 tokens billed as 72 seconds. The count includes the transcript, tool history and retrieved knowledge-base content.
- Burst costs more per minute. A call that starts above your concurrency limit pays $0.10 a minute extra “for the entire call duration”.
- Testing is not free. Retell’s testing page says tests bill “at the same rates as production” with “no free testing sandbox”.
The rule is headed “LLM Price Scaling” but written as billed duration, and does not say whether it scales the model line only or every per-minute line. For a 6,000-token agent (factor 1.5) at 5,000 minutes, the two readings give $489 or $701.50 on Stack A, $1,059 or $1,346.50 on Stack B, and $2,104 or $2,554 on Stack C. Ask Retell which reading applies before you build a long-prompt agent.
Retell beside a beta platform and a done-for-you service
Retell is the only one of these three that prints a per-minute rate you can multiply today.
| Option | How you buy it | Published price, as at 31 October 2026 |
|---|---|---|
| Zian (platform) | Partnership-application beta; each of its three listed plans, including Intel + Widgets (Free Forever), starts with a partnership application | No public price during beta |
| LeadsNow.ai (done-for-you) | Managed AI appointment setting and lead generation that runs on Zian | Pay-per-result; we do not print its price here |
| Retell AI (platform) | Self-serve pay-as-you-go from prepaid credits, or an Enterprise contract | $0.055/min Voice Infra plus LLM, voice and telephony; $0.093 to $0.340/min on our three stacks |
What Retell’s price list changed in the last 90 days
Retell’s changelog has five entries between 3 August and 31 October 2026: the Retell CLI (3 August), Retell Workflows (24 August), Conductor everywhere (15 September), a public model benchmark (29 September) and Conversation Memory (8 October). Only the last announces a price, “$0.005 per run” for saving and backfilling memory; Retell’s contact memory documentation adds that a backfill is charged “per conversation processed”, not per contact. None announces a per-minute rate change. The bigger change in the window sits in the documentation: the move to prepaid credits on 30 September 2026. Check the pricing page on the day you budget.
Retell AI pricing FAQ
How much does Retell AI cost per month?
There is no monthly plan fee on pay-as-you-go. As at 31 October 2026 you pay per minute: $0.055 for Retell’s voice engine plus the LLM, voice and telephony you choose, then monthly items such as $2 per Retell number. Our worked stacks come to $95, $181 or $342 for 1,000 US minutes, and $1,870 to $6,810 for 20,000.
How much is Retell AI per minute?
Retell quotes “$0.07-$0.31 / min for AI Voice Agents”. Built from its component table, our three US stacks cost $0.093, $0.179 and $0.340 a minute before monthly items. The LLM line alone runs from $0.0016 (GPT 5 nano) to $0.64 (GPT 6 Astra on Fast Tier), so the quoted range is not a ceiling.
Does Retell AI have a free trial?
Yes. New accounts get $10 in free trial credits, granted once per email address, with 20 free concurrent calls. At our leanest stack ($0.093 a minute) that is about 107 minutes. Testing bills at production rates, and buying a phone number needs a card on file.
Is Retell AI prepaid?
Yes, for pay-as-you-go. Retell’s billing documentation says usage is “deducted from your credit balance in real time”, and legacy monthly-billing workspaces were moved to credits on 30 September 2026. Phone numbers, extra knowledge bases, CPS and purchased concurrency still bill to your card at month end. Enterprise workspaces stay on monthly invoicing.
Can I get a refund on Retell credits?
No. Retell’s billing documentation says “Credits never expire, but they are non-refundable once purchased”, and its terms of service say all fees are non-refundable except as required by law or as the terms expressly provide. Spend the balance down before closing a workspace.
How much do extra concurrent calls cost on Retell?
Every workspace gets 20 concurrent calls free. Each extra one is $8 a month, charged when you raise the limit. Concurrency burst lets calls run above your limit for a $0.10 a minute surcharge on the whole call, up to the lower of 3 times your limit or your limit plus 300. Enterprise offers custom concurrency starting at 50+ calls.
Does Retell charge for voicemail, silence or transferred calls?
Retell charges only connected calls; on voicemail it bills only while the agent is on the line. Silence and hold time are billed. After a transfer the agent fee stops and only telephony continues. A call under 10 seconds with a dynamic opening message is billed as 10 seconds.
Where every figure on this page comes from
| Figure | Who published it | Link | Date read |
|---|---|---|---|
| Pay as you go from $0; “No minimums, no contracts, no surprises”; “$0.07-$0.31 / min”; “$0.002+ / msg”; Enterprise custom pricing | Retell AI | retellai.com/pricing | 31 October 2026 |
| Retell Voice Infra $0.055/min | Retell AI | retellai.com/pricing | 31 October 2026 |
| LLM prices per minute, $0.0016 (GPT 5 nano) to $0.32 standard, $0.64 Fast Tier, $0.384 Ultrafast; 36 rows; Recommended labels | Retell AI | retellai.com/pricing | 31 October 2026 |
| Speech-to-speech models $0.07 to $0.38/min; chat agent $0.0005 to $0.10 per AI message | Retell AI | retellai.com/pricing | 31 October 2026 |
| Voice prices: $0.015 (six families); ElevenLabs Flash $0.040, Multilingual v2 $0.06, v3 Conversational and v4 Turbo $0.08, v3 and v4 $0.10 | Retell AI | retellai.com/pricing | 31 October 2026 |
| Add-ons: Knowledge Base, Advanced Denoising, Safety Guardrails $0.005/min; PII Removal $0.01/min; Batch Call $0.005/dial; Branded Call $0.10/outbound call; AI QA 100 free minutes then $0.10/min; SMS $0.01/msg | Retell AI | retellai.com/pricing | 31 October 2026 |
| Monthly items: numbers $2.00; Retell SMS $20.00; concurrency 20 free then $8.00; knowledge bases 10 free then $8.00; verified number $10.00 (and “One-time $10 fee” note) | Retell AI | retellai.com/pricing | 31 October 2026 |
| Billed to the second; silence billed; transfers; connected calls and voicemail; 20 concurrent calls; Enterprise concurrency 50+ | Retell AI | retellai.com/pricing (FAQ) | 31 October 2026 |
| Country telephony rates: US $0.015, US toll-free $0.06, Canada $0.03, Australia $0.10, Philippines $0.80, Telnyx US $0.03 | Retell AI | International calling | 31 October 2026 |
| Numbers US/Canada $2/month, US toll-free $5/month, toll-free inbound $0.06/min | Retell AI | Purchase phone number | 31 October 2026 |
| Prepaid credits; card items; $10 trial once per email; credits non-refundable; auto recharge $10–$1,000 and $20–$5,000; budget $10–$100,000; Enterprise monthly invoicing | Retell AI | Billing: credits, auto recharge, and invoices | 31 October 2026 |
| Legacy monthly billing moved to credits on 30 September 2026; SMS subscriptions bill monthly | Retell AI | Legacy monthly billing notice | 31 October 2026 |
| 10-second minimum; 4,000-token scaling formula; 60-second, 4,800-token example billed as 72 seconds | Retell AI | Exceptions to Our Per-Minute Pricing | 31 October 2026 |
| 20 concurrent calls per workspace; burst $0.10/min on the whole call; lower of 3 times or limit plus 300; Dialer never uses burst | Retell AI | Understand concurrency & limits | 31 October 2026 |
| Extra concurrency $8.00 charged immediately; local $2 and toll-free $5 numbers; daily spam check $2 per number per month | Retell AI | Caller ID and pacing | 31 October 2026 |
| Batch calls deprecated in favour of the Dialer | Retell AI | Create and schedule batch calls | 31 October 2026 |
| DNC screening “billed per number checked” | Retell AI | Compliance checks | 31 October 2026 |
| SMS $20/month per number plus $0.01 per SMS; brand $4 or $45; campaign $15 non-refundable; US only, not toll-free, not Telnyx | Retell AI | Enable SMS | 31 October 2026 |
| Knowledge base $0.005/min however many are used; first 10 free, then $8/month | Retell AI | Knowledge base setup | 31 October 2026 |
| AI QA first 100 minutes free per workspace, then $0.10/min | Retell AI | AI QA | 31 October 2026 |
| Testing billed at production rates, no free sandbox; Conductor free to 30 credits per user and 200 per workspace a day, then $0.20 per credit if pay-as-you-go is on | Retell AI | Testing pricing | 31 October 2026 |
| Fast Tier price uplift varies by model; fast fallback can bill both models | Retell AI | LLM options | 31 October 2026 |
| Branded call billing runs to cycle end on stopping; verified number stops immediately; both US only | Retell AI | Stop caller ID service; Apply for branded call | 31 October 2026 |
| Fees non-refundable; renewal unless items removed; 14-day invoice, 7-day grace, late fee lesser of 10% or legal maximum, suspension after the grace period; rate changes from next cycle | Retell AI | Terms of Service (last updated 25 October 2026) | 31 October 2026 |
| Changelog entries 3 August to 8 October 2026; Conversation Memory $0.005 per run; backfill $0.005 per conversation processed | Retell AI | retellai.com/changelog; Conversation Memory; Contact memory | 31 October 2026 |
| Worked bills ($95 to $6,810), per-minute stack rates, outbound and long-prompt variants | Zian AI calculation from the Retell rates above | This page, “Retell AI cost worked at 1,000, 5,000 and 20,000 minutes” | 31 October 2026 |
| 50,769+ qualified sales appointments set | Zian AI | zian.ai | 31 October 2026 |
| LeadsNow.ai pay-per-result; leads answered in under 60 seconds, 24/7 | LeadsNow.ai | leadsnow.ai | 31 October 2026 |
Ready to see an AI receptionist and outbound agent on one platform? Zian answers inbound calls 24/7, books appointments, handles voicemail on calls and follows up by voice, SMS, email, WhatsApp, Instagram DMs and Facebook Messenger. It is in partnership-application beta. Apply For Partnership