ElevenLabs Agents Pricing: $0.08 a Minute, Reception From $29

October 21, 2026. ElevenLabs charges $0.08 for every minute of an AI agent call on every self-serve plan, from the free tier to the $990 Business plan, according to its ElevenAgents pricing page, read today. The plan sets how many minutes you prepay and how many calls can run at once, not the rate. Calls above your concurrency limit cost $0.16 a minute, text messages $0.003 each, and the language model is billed on top at the provider's own price. For owners who want the receptionist without the build, ElevenLabs also sells Reception, a packaged AI receptionist, from $29 a month for 75 minutes. Here is every rate, what each plan really buys, and what 1,000 minutes cost built or bought.

ElevenAgents rates, read October 21

1. Call minutes. $0.08 for each additional minute on Free and pay as you go, Starter, Creator, Pro, Scale and Business alike. ElevenLabs cut the Starter rate from $0.10 to $0.08 in a pricing post dated May 7 and last updated October 19, which also introduced pay as you go.
2. Plans. Free is $0 with 15 minutes and 4 concurrent calls. Starter is $6 with 75 minutes and 6 calls. Creator is $22 with 275 minutes and 10 calls, at $11 for the first month. Pro is $99 with 1,238 minutes and 20 calls. Scale is $299 with 3,738 minutes, 30 calls and 3 seats. Business is $990 with 12,375 minutes, 40 calls, 10 seats and 10 professional voice clones. Enterprise is custom and adds BAAs for HIPAA customers.
3. Burst. With burst switched on, an agent can take up to three times your concurrency limit or 300 calls, whichever is lower, and the extra calls bill at double the rate, $0.16 a minute. The burst pricing docs add that burst calls get lower priority for speech processing, and that calls past the burst limit are rejected unless call queueing is on.
4. Text and silence. Text messages are $0.003 each. Voice calls get a 95 percent discount for silences longer than 10 seconds, and a call is measured on connection time, from the moment it starts until it ends or the window closes, according to the help center.
5. The language model. ElevenLabs passes third-party model costs through at the provider's published rate with no markup, its models documentation says, except that select Gemini and Claude models bill US and EU traffic at Vertex AI regional rates, a 10 percent increase.
6. Telephony. ElevenLabs adds no telephony fee. You connect Twilio or a SIP trunk, and that carrier bills you directly.

What a plan really buys

Divide each plan's price by its minutes: $6 for 75, $22 for 275, $99 for 1,238, $299 for 3,738 and $990 for 12,375 all come to $0.08 a minute, the same as an additional minute. A bigger plan does not make a minute cheaper. It buys concurrency, from 4 calls at once on Free to 40 on Business, plus seats and voice clones. The one discount is yearly billing, which gives two months free and brings the paid plans to $5, $18.33, $82.50, $249.17 and $825 a month, about $0.067 a minute.

The plan also decides what happens when the minutes run out. The plan cards list additional minutes as a feature from Creator up, and the help center says you need pay-as-you-go credits, or usage-based billing on a legacy subscription, before you can exceed your credit quota. Pay as you go itself has no plan fee: the pricing page calculator prices 270 minutes on GPT-5.4 Mini at $23.41, which is $0.08 of voice plus $0.0067 of model per minute. In the price table it shares the free tier's 4 concurrent calls.

So the plan question is really a concurrency question. A one-van plumber who rarely sees two calls at once can stay on pay as you go. An HVAC company whose phones ring eight at once on the first morning of a heat wave needs at least Creator's 10 lines, or burst at double the rate.

The language model is the line that moves

The calculator's per-minute model costs, read October 21:

1. GPT-5 Nano $0.0005, Gemini 2.5 Flash Lite $0.0008, GPT-4o Mini $0.0012, GPT-5 Mini $0.0024, and Gemini 2.5 Flash $0.0029, the calculator's default.
2. GPT-5.4 Mini $0.0067, Claude Haiku 4.5 $0.0095 and Gemini 3.5 Flash $0.0147.
3. Claude Sonnet 4.6 $0.0284, GPT-5.5 $0.0446 and Claude Opus 4.7 $0.0474.

Over 1,000 minutes that is a range from $0.50 to $47.40, so the model can add up to 59 percent to the $80 of voice minutes. OpenAI retires GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini from its API on January 14, 2027, according to its deprecations page, and names GPT-6 Luna as the replacement, a model ElevenLabs' models page already lists. Our GPT-4o API shutdown brief covers the cut-off.

Reception: the $29 receptionist ElevenLabs sells itself

ElevenLabs launched Reception, a self-serve AI receptionist for small businesses built on ElevenAgents, on September 16. It answers calls, books jobs into the business calendar and sends call summaries afterwards. Inc. reported the launch the next day with the same three prices that the Reception plans and pricing docs list today:

1. Basic, $29 a month. 75 credits, 1 phone number, 1 concurrent call. Extra phone minutes $0.45.
2. Plus, $79 a month. 275 credits, 3 numbers, 3 concurrent calls. Extra minutes $0.38.
3. Premium, $199 a month. 1,000 credits, 5 numbers, 10 concurrent calls. Extra minutes $0.30.

A phone minute uses 1 credit, a web chat minute 0.5 and a text 0.1, according to the usage and limits page. Unused credits do not roll over, and annual billing brings the plans to $24, $66 and $166 a month. The 14-day trial carries 30 credits and no overage: when they run out, the receptionist stops answering until you upgrade. Basic's included minutes work out to about $0.39 each, nearly five times the ElevenAgents rate, and that premium pays for booking, call summaries and the calendar link arriving ready-made.

What 1,000 minutes cost, built or bought

Take 1,000 inbound phone minutes in a month at the list prices above:

1. ElevenAgents on pay as you go with GPT-5 Nano: $80.50. With GPT-5 Mini, $82.40. With Claude Opus 4.7, $127.40.
2. Reception Premium: $199. All 1,000 minutes sit inside the plan.
3. Reception Plus: $354.50. $79 plus 725 extra minutes at $0.38.
4. Reception Basic: $445.25. $29 plus 925 extra minutes at $0.45, on a line that takes one call at a time.

The ElevenAgents rows leave out the carrier. On Twilio's US price list, receiving a call on a local number costs $0.0085 a minute and the number $1.15 a month, so about $9.65 more on the simplest wiring; our Twilio Voice pricing breakdown prices the other ways to connect. They also leave out the build: the prompt, the booking integration, transfers and testing. And the cheapest Reception plan is the most expensive way to buy 1,000 minutes, because Basic is priced for a business that takes a handful of calls a day.

What it means for operators

For a home services company the raw minute is no longer the decision. ElevenAgents' $0.08 covers speech recognition and the voice but not the model, while Deepgram's $0.075 Standard tier in our Deepgram pricing brief includes a managed model, so the meters sit within cents of each other. What costs money is everything around the minute: concurrency for the busy hour, the job-booking integration and transfers to an on-call technician. Reception sells booking, texts and transfer rules ready-made at a premium per minute, with few lines on its cheaper plans; ElevenAgents sells the meter and leaves the rest to whoever builds it.

1. If you already pay ElevenLabs, check your pricing version. The pricing post says existing customers move by opening Manage subscription and clicking Switch to new pricing; if the button is missing, you are already on it.
2. Pick the model before the plan. GPT-5 Mini or Gemini 2.5 Flash keeps the model under $3 per 1,000 minutes, and neither is one of the four OpenAI models retiring in January.
3. Size concurrency to your worst hour. Turn on burst for storm days knowing those minutes cost double, and switch on call queueing so callers past the limit wait instead of being rejected.
4. Buy pay-as-you-go credits so an agent that uses up its included minutes in the third week of a busy month keeps answering.
5. Trial Reception on a forwarded line before you commit; the 30 trial credits are a hard stop.

Our AI receptionist cost guide compares the monthly bill across 11 vendors, the best AI voice agents comparison covers transfers and features, and for HVAC, see which AI answering services book the job. For text to speech on the API, see our ElevenLabs v4 Turbo pricing brief. If you would rather have the agent built, connected to your dispatch software and tested on your own call types, see our AI call answering for home services, or hire an AI engineer for a custom build.