ElevenLabs Pricing Per Minute: $0.04 to $0.20, Plan by Plan

Short answer: ElevenLabs text-to-speech costs about $0.08 per minute on the API for its v2, v3 and v4 voices and about $0.04 per minute for Flash, Turbo and v3 Conversational. Inside the web app you pay in credits instead, which works out to $0.165 to $0.20 per minute depending on your plan. Voice agents are billed separately at $0.08 per call minute beyond your plan’s included minutes, plus the LLM and your phone carrier.

The rest of this guide shows the plan table, how to convert credits to minutes, the API and Agents rates, and two worked examples: a 10-minute podcast and 1,000 agent minutes. Prices were checked on ElevenLabs’ own pricing pages on 28 October 2026, are in USD and exclude tax. ElevenLabs changes its pricing often, so confirm before you commit.

The one conversion you need: about 1,000 characters per minute

ElevenLabs does not sell “minutes” of speech directly. It sells characters, either as credits in the app or as dollars per 1,000 characters on the API. The bridge is simple:

• ElevenLabs’ API pricing page prices v4, v3 and Multilingual v2 at $0.08 per 1K characters and labels that “~$0.08/min”. So ElevenLabs itself treats a minute of audio as roughly 1,000 characters.
• On the plans page, Multilingual v2 text-to-speech costs 1 credit per character, and the page lists each plan’s included minutes at about one minute per 1,000 credits.

Your real number depends on how fast the voice speaks and how much punctuation and pausing your script has. Use 1,000 characters per minute for budgeting, then check your first few generations.

ElevenLabs plans and credits (as of October 2026)

These are the self-serve plans on elevenlabs.io/pricing, checked 28 October 2026.

How to read it:

• Effective $ per minute is simply the plan price divided by its included minutes, for example $22 ÷ 121 = $0.182. That is what you pay per minute if you use the whole allowance.
• Annual billing costs ten times the monthly price, so you get two months free. Creator billed yearly is $220, or $18.33 a month.
• Rollover: unused credits roll over for up to two months, capped at three times your monthly quota.
• Seats: Scale includes 3 seats and Business 10.
• Promotions: at the time of checking, ElevenLabs showed a first-month promo on Starter ($1) and Creator ($11) running until 1 November. Treat it as a one-off.

Credits are not only for text-to-speech

The same credit pool pays for other tools, at very different rates, so a “600-minute” plan can disappear fast if you mix them:

Dubbing is the one to watch. At the top of that range, one minute of dubbing uses as many credits as ten minutes of narration.

ElevenLabs API pricing per minute

The API pricing page bills in US dollars rather than credits. Your plan fee covers usage up to the same dollar amount, and the included characters scale with it.

Included API characters for the v4 voice class, per the same page: Free 10,000 (about 10 minutes), Starter 75,000, Creator 275,000, Pro 1,238,000, Scale 3,738,000 and Business 12,375,000. Turbo-class voices get double those amounts because they cost half as much.

Notice the pattern: $6 ÷ $0.08 per 1K = 75,000 characters, and $22 ÷ $0.08 = 275,000. On the API the plan fee is essentially a prepaid minimum at $0.08 per 1K characters, which is why the per-minute API price is roughly half the app’s effective rate.

The Business plan also advertises low-latency TTS “as low as 5¢/min” on the plans page. That is a floor for high-volume customers, not the default rate.

ElevenLabs Agents pricing per minute

Voice agents (the conversational product) are billed by call minutes, not characters. From elevenlabs.io/pricing/agents:

• Additional minutes: $0.08 each on every plan.
• Burst minutes: $0.16 each, for calls above your concurrency limit (up to 3x concurrency).
• SMS: $0.003 per message.
• LLM: billed separately and deducted from your credits.
• Telephony: ElevenLabs adds no fee. Twilio or your SIP provider bills you directly.

Again, the included minutes equal the plan price at $0.08 a minute ($22 ÷ $0.08 = 275). So the true agent rate is a flat $0.08 per minute, and the plan you pick mainly sets your minimum spend and your concurrency.

For how ElevenLabs Agents compare with Vapi, Retell and Bland once you add every layer of the stack, see our AI voice agent cost per minute breakdown.

Worked example 1: a 10-minute podcast episode

Assume a solo or two-host script of about 10,000 characters (10 minutes at ~1,000 characters per minute). Multi-speaker does not change the maths; you pay per character either way.

Option A: web app on Creator ($22/month, 121,000 credits)

• One clean pass: 10,000 characters × 1 credit = 10,000 credits.
• Allow 30% for retakes and tweaks: 10,000 × 1.3 = 13,000 credits per episode.
• Episodes per month: 121,000 ÷ 13,000 = 9 episodes (9.3, rounded down).
• Cost per episode: $22 ÷ 121,000 × 13,000 = $2.36 (or $1.82 with no retakes).

Option B: API with a v4 or Multilingual v2 voice

• 10,000 characters × $0.08 per 1K = $0.80 per clean episode.
• With 30% retakes: 13,000 × $0.08 per 1K = $1.04.
• With a Flash/Turbo voice at $0.04 per 1K: $0.40 to $0.52.

Option C: Starter ($6, 30,000 credits) gives you about two episodes a month with retakes (30,000 ÷ 13,000 = 2.3), at $2.60 each if you use the whole allowance.

The 30% retake allowance is our assumption, not an ElevenLabs figure. Track your own ratio for a month; heavy editors can easily double their character use.

A weekly show (about four or five episodes a month) fits comfortably in Creator. A daily show needs Pro: 22 episodes × 13,000 = 286,000 credits, well under Pro’s 600,000.

Worked example 2: 1,000 voice-agent minutes a month

Assume a US inbound support line, 1,000 call minutes a month, on a Twilio local number.

ElevenLabs fees

Starter and Creator land on the same $80 because included minutes are priced at $0.08 too. Pick Creator if you need more than 6 simultaneous calls; Pro only makes sense here if you need 20 concurrent calls or expect to grow past 1,237 minutes.

Add the LLM and the phone line

• LLM: ElevenLabs bills it separately and the rate depends on the model. As a reference range, Vapi’s public calculator shows OpenAI models at $0.0077 to $0.0452 per minute at 1,000 minutes. That is $7.70 to $45.20 a month.
• Telephony: Twilio charges $0.0085 per minute to receive calls on a US local number plus $1.15 a month for the number (Twilio US). That is $8.50 + $1.15 = $9.65.

Total: $80 + $7.70 to $45.20 + $9.65 = $97.35 to $134.85 a month, or about $0.10 to $0.13 per minute all-in.

If calls spike above your concurrency limit, those minutes bill at $0.16 instead of $0.08, so size your plan’s concurrency to your busiest hour.

For Australia: receiving calls on an Australian local Twilio number costs $0.0100 per minute plus $3.00 a month (Twilio AU), so telephony is $13 instead of $9.65. Canadian local numbers are priced the same as US ones.

Prices in CAD and AUD

At the 29 October 2026 rate from ExchangeRate-API (1 USD = 1.40 CAD = 1.42 AUD), before tax:

Your card issuer’s exchange rate and fees will differ slightly.

Tax: ElevenLabs says its prices exclude taxes. Australia’s GST is 10% and applies to imported digital services sold to Australian consumers; sales to Australian GST-registered businesses are treated differently (ATO). Canada has applied GST/HST rules to foreign digital-economy businesses since July 1, 2021 (CRA). If you run a US company from Canada or Australia, the tax on your invoice depends on the billing entity and address you use. See our guides for Canadians with a US LLC and Australian founders with a US company. This is general information, not tax advice.

Which plan should you pick?

Two rules of thumb. First, if you write code, the API is the cheapest way to buy ElevenLabs speech per minute. Second, for agents, the plan barely changes the per-minute price; it sets concurrency and your monthly minimum, so buy for your peak call volume, not your average.

Common mistakes that inflate your bill

• Mixing dubbing into a narration plan. At 2,000 to 10,000 credits per minute, a few minutes of dubbing can eat a month of narration credits.
• Ignoring retakes. Budget at least 20 to 30% extra characters if you regenerate lines.
• Forgetting the LLM and carrier on agents. The $0.08 per minute ElevenLabs fee can be as little as 60% of the real cost.
• Underbuying concurrency. Burst minutes cost double.
• Letting credits expire. Rollover lasts two months and caps at three times your quota.

Bottom line

For narration, ElevenLabs costs about $0.08 per minute on the API and $0.165 to $0.20 per minute through app credits. For voice agents, budget $0.08 per minute for ElevenLabs plus roughly $0.02 to $0.05 for the LLM and phone line. If you are comparing it with other AI tools on your stack, our Claude Code vs Codex vs Cursor pricing guide uses the same sourced, dated approach, and the AI tools hub collects the rest. Building a text agent instead? See n8n AI agent cost per run.