Tools Web

Calculators

TTS API Cost Comparison

Enter monthly audio minutes and characters spoken per minute. The comparison converts OpenAI per-character, ElevenLabs per-1K-character, Gemini per-10-second and Cartesia credit pricing into one $ per 1M characters and $ per audio minute, then prices the same monthly sample. Estimates only; vendor pricing pages prevail.

Math runs in your browser. Nothing is uploaded.

Same sample for every vendor
—

VendorModel / plan$ / 1M chars$ / audio minAt your volume / moNote

Assumptions: OpenAI tts-1 $15 and tts-1-hd $30 per 1M characters (gpt-4o-mini-tts is billed per audio token with no published per-minute rate, so it is left out). ElevenLabs API pay-as-you-go per 1,000 characters: Flash / Turbo and v4 Turbo $0.04, v3 and Multilingual v2 $0.08. Gemini TTS uses Google's official per-10-second audio equivalents (Standard tier; 3.8 Flash / Flash-Lite at promo rates through 2026-12-31, doubling from 2027-01-01) plus text input at ~4 characters per token. Cartesia Sonic is 1 credit per character; "included" rows are plan fee ÷ credits, the last row is the cheapest plan fit with optional overage ($65 / $45 / $38 per 1M). Characters per minute (default 800) converts between units. STT, LLM, telephony and voice-agent fees are not included; voice quality and latency differ — test your own script.

Estimates only. List prices as of 2026-11-01 (Asia/Shanghai); vendor pricing pages prevail. Sources: OpenAI pricing · ElevenLabs API pricing · Gemini API pricing · Cartesia pricing

Voice agent platforms (full call minute)

TTS is only one slice of an AI phone-agent minute. Platforms bundle STT + LLM + TTS + telephony: Vapi charges $0.05/min hosting plus pass-through model costs (≈ $0.08–$0.13/min all-in on its own example), Retell bills components (default $0.11/min = GPT 5 $0.04 + infra $0.055 + voice $0.015). Bland bundles LLM + STT + TTS at $0.14/min (Start) or $0.12/min + $299/mo (Build). Speech-to-speech APIs (OpenAI Realtime, Gemini Live) price audio tokens directly.

Voice AI Agent Platform Cost Comparison · Vapi Cost Calculator · Retell AI Cost Calculator · Bland AI Cost Calculator · OpenAI Realtime Voice Cost Calculator · Gemini Live Voice Cost Calculator · Deepgram API Cost Calculator · ElevenLabs TTS Cost Calculator · Cartesia TTS Cost Calculator

Vapi Cost Calculator · Retell AI Cost Calculator · Cartesia TTS Cost Calculator · ElevenLabs TTS Cost Calculator · OpenAI TTS Cost Calculator · Gemini TTS Cost Calculator · OpenAI Realtime Voice Cost Calculator · Gemini Live Voice Cost Calculator

More in this group

How it works

Which TTS API is cheapest per 1M characters?
At 800 characters per minute: Gemini 3.8 Flash-Lite TTS ≈ $11.38 (promo through 2026-12-31), OpenAI tts-1 $15, Gemini 3.8 Flash ≈ $17 (promo), OpenAI tts-1-hd $30, Cartesia Scale ≈ $37.38 (included credits), Gemini 3.1 Flash TTS preview ≈ $37.75, ElevenLabs Flash/Turbo $40, Eleven v3 / Multilingual v2 $80. Gemini bills by audio length, so its $ per 1M characters depends on speaking rate.
What does 10,000 audio minutes a month cost?
At 800 characters per minute (8M characters): OpenAI tts-1 $120, tts-1-hd $240, ElevenLabs Flash/Turbo $320, Eleven v3 $640, Gemini 3.8 Flash TTS ≈ $136 (promo), Flash-Lite ≈ $91, Gemini 3.1 Flash preview ≈ $302, Cartesia Scale $299 (8M credits fit exactly).
Why compare per minute and per character?
Vendors quote different units: characters, credits per character, or seconds of audio. Normalizing to $ per audio minute shows what a voice-agent or IVR minute costs; $ per 1M characters is easier for scripts and articles. Speaking rate (750–900 characters per minute) shifts the result.
Is voice quality or latency included?
No. This page compares list price only. Latency, voice cloning, languages and quality differ; test your own script before choosing. STT, LLM and telephony add to a full voice-agent minute.
Is any data uploaded?
No. These calculators and QR codes run in your browser.