InputsUSD · pay-as-you-go list prices
Unit
Quality

Most vendors bill per input character; Gemini TTS and gpt-4o-mini-tts bill by audio time, converted here through the pace field and marked ≈. Only free tiers that renew every month enter the math — Amazon Polly's 12-month trial quotas stay out so long-run costs aren't understated.

Ranking150K chars/mo
Cheapest at this volume
$0/mo
Google Cloud Standard voices

Google Cloud Standard voices covers this volume inside its monthly free tier — $0/mo.

  • Google Cloud Standard voices$4.00/1M free tier $0/mo
  • Google Cloud WaveNet voices$4.00/1M free tier $0/mo
  • Azure AI Speech Neural / Neural HD Flash$15.00/1M free tier $0/mo
  • Google Cloud Neural2 voices$16.00/1M free tier $0/mo
  • Google Cloud Chirp 3: HD voices$30.00/1M free tier $0/mo
  • Google Cloud Studio voices$160/1M free tier $0/mo
  • Amazon Polly Standard voices$4.00/1M budget $0.600/mo
  • Deepgram Aura-1$15.00/1M standard $2.25/mo

List prices verified on vendor pages 2026-08-13. ≈ rows are time-billed models converted at your pace setting; gpt-4o-mini-tts additionally rests on OpenAI's own "~$0.015/min" estimate. Volume discounts, enterprise contracts, and regional prices not modeled.

Embed this calculator ↓
Start from your profile

Six usage profiles

Click one to load it. Verdicts are computed from the current rate sheet — they move when a vendor reprices.

All rates

Your volume against every rate

The scenario configured above, priced on all 22 verified rates. Rows keep the default-scenario order; the rank and monthly columns recompute live as you edit inputs.

# Model Meter $ / 1M chars Free tier Your volume Src
1 Google Cloud Standard voices budget per character $4.00 4M/mo $0
2 Google Cloud WaveNet voices budget Cut from $16 to $4 per 1M chars per character $4.00 4M/mo $0
3 Azure AI Speech Neural / Neural HD Flash standard East US list price — varies by region per character $15.00 500K/mo $0
4 Google Cloud Neural2 voices standard per character $16.00 1M/mo $0
5 Google Cloud Chirp 3: HD voices premium per character $30.00 1M/mo $0
6 Google Cloud Studio voices premium Narration-grade studio voices per character $160 1M/mo $0
7 Amazon Polly Standard voices budget per character $4.00 12-mo trial $0.600
8 Deepgram Aura-1 standard Growth prepay drops it to $13.50 per character $15.00 $2.25
9 OpenAI tts-1 standard Latency-optimized per character $15.00 $2.25
10 OpenAI gpt-4o-mini-tts standard Steerable voices (“vibes”) per minute ≈ $15.11 ≈$2.27
11 Google Gemini 2.5 Flash TTS standard Prompt-steerable LLM voices per minute ≈ $15.13 ≈$2.27
12 Amazon Polly Neural voices standard per character $16.00 12-mo trial $2.40
13 Deepgram Aura-2 premium Growth prepay drops it to $27 per character $30.00 $4.50
14 OpenAI tts-1-hd premium Quality-optimized per character $30.00 $4.50
15 Amazon Polly Generative voices premium per character $30.00 12-mo trial $4.50
16 Google Gemini 2.5 Pro TTS premium per minute ≈ $30.25 ≈$4.54
17 Google Gemini 3.1 Flash TTS premium Preview pricing per minute ≈ $30.25 ≈$4.54
18 ElevenLabs Flash / Turbo premium ~75 ms latency per character $50.00 20K/mo $6.50
19 MiniMax speech-2.8-turbo standard per character $60.00 $9.00
20 ElevenLabs Multilingual v2 / v3 premium Flagship quality per character $100 10K/mo $14.00
21 MiniMax speech-2.8-hd premium per character $100 $15.00
22 Amazon Polly Long-form voices premium per character $100 12-mo trial $15.00

$ / 1M chars for time-billed models (marked ≈) is derived at the current pace setting — 1000 characters per minute by default. Free tier column shows monthly-recurring quotas; Amazon Polly's 12-month trial quotas are noted but never enter the math. Src links open the vendor pricing page each rate was verified on.

ElevenLabs pricing

The ElevenLabs plan sheet

Since the 2026 repricing the per-character rate is identical on every tier — $0.05/1K on Flash, $0.10/1K on Multilingual — so each plan includes exactly its face value of synthesis. What upgrading buys is the commercial license, voice cloning, and concurrency. Verified 2026-08-13 on the ElevenLabs pricing page.

Plan $ / mo Multilingual chars Flash chars ≈ audio
Free $0 10K 20K 10–20 min
Starter $6.00 60K 120K 60–120 min
CreatorFirst month $11 $22.00 220K 440K 220–440 min
Pro $99.00 990K 2.0M 990–1,980 min
Scale $299 3.0M 6.0M 2,990–5,980 min
Business $990 9.9M 19.8M 9,900–19,800 min

Included characters differ by model because Flash bills half of Multilingual — the audio column shows the range (Multilingual minutes – Flash minutes at ~1K chars/min). Overage bills at the same per-1K rate as the plan itself.

How it works

The math behind the ranking

Every rate is a US pay-as-you-go list price read off the vendor's own pricing page, stamped with its verification date. The engine normalizes three metering schemes into one monthly number:

/* per-character vendors (most rows) */
billable  = max(0, chars − monthly_free_tier)
monthly   = billable × rate_per_1M ÷ 1M

/* time-billed vendors (Gemini TTS, gpt-4o-mini-tts) — marked ≈ */
minutes   = chars ÷ pace                 /* pace: 1,000 chars/min default */
monthly   = minutes × rate_per_minute
          + (chars ÷ chars_per_token) × text_input_rate ÷ 1M

Gemini TTS's per-minute rate is exact — Google publishes the conversion (25 audio tokens per second). gpt-4o-mini-tts rests on OpenAI's own "~$0.015 per minute" launch estimate, so its rows carry the ≈ mark twice over. Not modeled: volume discounts, enterprise contracts, regional price variation (Azure's figure is the East US list), and Amazon Polly's 12-month trial quotas.

Full methodology lives at /methodology/. Found a stale rate? Report it.

Worked examples

Three volumes, real numbers

Computed from the current rate sheet — recalculated on every data refresh.

Example 01 · audiobook
Same book, $0 to $54.00
A 550K-character manuscript (~9 hours of narration) through three premium voices. The text is identical; only the rate changes.
Volume550K chars/mo
Google Chirp 3 HD$0 (free tier)
OpenAI tts-1-hd$16.50/mo
ElevenLabs Multilingual$54.00/mo
$0–$54.00
voice choice is the entire budget
Example 02 · voice agent
2M characters still rides a free tier
An IVR generating 2M characters of prompts a month — half of Google's 4M-character Standard quota, so the bill is zero until quality requirements raise it.
Volume2M chars/mo
Google Standard$0 (free tier)
Deepgram Aura-2$60.00/mo
ElevenLabs Flash$99.00/mo
$0
recurring free tiers are real infrastructure
Example 03 · at scale
20M characters: rate is everything
A read-aloud feature pushing 20M characters (~333 hours) a month. Past every free tier, the per-million rate is the whole story.
Volume20M chars/mo
Google Standard$64.00/mo
OpenAI tts-1$300/mo
ElevenLabs Flash$999/mo
$935
monthly gap between premium and budget at this volume
Characters, tokens, minutes

Why TTS pricing resists comparison

01
Characters include the markup

Per-character vendors count everything you send: spaces, newlines, and SSML tags. Google bills every SSML tag except <mark> — a heavily annotated script costs measurably more than its visible text. Budget on the string you POST, not the words you hear.

02
There are two Googles on this page

Classic Cloud TTS bills per character ($4–$160 per 1M by voice tier). Gemini TTS is a different product billed per token, with a published conversion — audio is 25 tokens per second — which is what lets us print an exact $0.015/min for Gemini 2.5 Flash TTS. Same company, incomparable meters.

03
A minute of speech is 750–1,000 characters

Vendors approximate 1K characters ≈ 1 minute (that's ElevenLabs' own rule of thumb); measured speech at a natural pace runs closer to 750–950. The pace field converts minutes to characters — nudge it down for slow narration and time-billed rows will look relatively better.

04
Free tiers come in three species

Monthly-recurring quotas (Google's 4M standard / 1M premium characters, Azure's 0.5M) renew forever and are real money at small volumes. Plan-included characters (ElevenLabs) renew but expire monthly. Trial quotas (Amazon Polly's first-12-months allotments) end — this calculator keeps them out of the math so year-two costs aren't understated.

05
ElevenLabs plans buy face value, not cheaper characters

Since the 2026 repricing, the per-1K rate is identical on every ElevenLabs tier — $22 Creator includes exactly $22 worth of Multilingual synthesis. Upgrading buys the commercial license, voice cloning, and concurrency, not a volume discount. The only real discount lever is the model: Flash bills half of Multilingual.

TTS pricing — answered.

The questions this calculator exists for, answered with the current numbers. Every figure recomputes when a vendor reprices.

Q · 01 How much does ElevenLabs cost? +
Two prices, flat across all plans since the 2026 repricing: $0.05 per 1K characters for Flash/Turbo and $0.10 per 1K for Multilingual v2/v3 (about $50 and $100 per million). Subscriptions bundle exactly their face value of usage — Starter $6 includes 60K Multilingual characters, Creator $22 includes 220K, Pro $99 includes 990K — plus the commercial license and cloning features. The plan table below has the full sheet.
Q · 02 Is ElevenLabs cheaper than OpenAI TTS? +
No — it's 6.7× OpenAI's price at flagship quality. ElevenLabs Multilingual bills $100/1M characters vs $15/1M for OpenAI tts-1 ($30 for tts-1-hd). Even ElevenLabs' cheaper Flash model ($50/1M) runs 3.3× tts-1. What you're paying for is voice quality, cloning, and 32-language coverage — run your volume through the calculator and decide if the gap is worth it at your scale.
Q · 03 What is the cheapest text-to-speech API? +
At list price: Google Cloud Standard/WaveNet and Amazon Polly Standard, both $4 per 1M characters — and Google's 4M-character monthly free tier means many workloads pay $0 indefinitely. Cheapest premium-tier voices cluster at $30/1M: OpenAI tts-1-hd, Google Chirp 3 HD, Amazon Polly Generative, Deepgram Aura-2. The spread from cheapest to priciest rate on this page is 40× — model choice dwarfs every other optimization.
Q · 04 How many characters is one minute of speech? +
Roughly 750–1,000 characters per minute at a natural reading pace (~150 words/min × ~5–6 characters per word). Vendors round to 1K ≈ 1 minute — ElevenLabs prints "$0.05/min" next to its $0.05/1K rate. This calculator uses 1,000 by default and lets you adjust the pace; 1M characters is about 1,000 minutes (~16.7 hours) of audio.
Q · 05 How does OpenAI TTS pricing work? +
Three models, two meters. tts-1 ($15/1M characters) and tts-1-hd ($30/1M) bill per input character. gpt-4o-mini-tts bills per token — $0.60/1M text input tokens plus $12/1M audio output tokens, which OpenAI itself estimates at ~$0.015 per minute of audio. That estimate makes it the cheapest OpenAI voice for most workloads, and it takes style instructions ("speak like a sympathetic customer service agent") the tts-1 family ignores.
Q · 06 Which TTS APIs have free tiers? +
Recurring monthly: Google Cloud (4M characters/mo on Standard/WaveNet, 1M on Neural2, Chirp 3 HD and Studio), Azure Speech (0.5M/mo), and ElevenLabs' free plan (10K Multilingual or 20K Flash characters). Amazon Polly's free tier is a 12-month trial, not a recurring quota — this calculator excludes it from the math so long-run costs stay honest. For hobby volumes the recurring tiers cover everything: the ranking's top rows read $0.
Q · 07 How is Gemini TTS priced? +
Per token, with a published conversion: audio output is 25 tokens per second. Gemini 2.5 Flash TTS bills $0.50/1M text input tokens + $10/1M audio output tokens — exactly $0.015 per minute of audio. Gemini 2.5 Pro TTS and the Gemini 3.1 Flash TTS preview run $1 + $20/1M, or $0.03/min. These are prompt-steerable LLM voices, a different product from Google's classic per-character Cloud TTS tiers on the same pricing page.
Q · 08 What does a full audiobook cost to synthesize? +
A 550K-character manuscript (~9 hours of audio) ranges from $0 to $54.00 depending on the voice: Google Chirp 3 HD rides its 1M-character free tier to $0, OpenAI tts-1-hd bills $16.50, ElevenLabs Multilingual $54.00. Same text, 40× rate spread — which is the point of this page.
Embed

Put this calculator on your site

Free to embed — your readers always see current rates. The widget links back here for the full rate table.

<iframe
  src="https://aicost.tools/calculator/text-to-speech/embed/"
  width="100%" height="760" loading="lazy"
  style="border:1px solid #3d3a32"
  title="Text-to-Speech Cost Calculator by AI//COST">
</iframe>
More calculators

The cost toolbox

Every tool runs on the same live-pricing backbone. See the full index →

One weekly digest. Zero noise.

One weekly digest · No spam, ever · Unsubscribe in one click