All voice AI prices

Every row links to the vendor page it came from. Click a column header to sort. Per-minute figures for LLM and TTS depend on the assumptions below.

Conversion assumptions
AssumptionDefaultWhy
turns_per_min4LLM calls per call-minute (one per caller utterance). Adjustable in the calculator.
input_tokens_per_turn2000System prompt + tools + running conversation history re-sent on every turn. 2,000 is a short prompt; long prompts and late-call history push this well above 5,000.
output_tokens_per_turn60~60 tokens is roughly 45 spoken words, one or two sentences.
cached_input_share0Share of input tokens billed at the vendor's cached-input rate. Default 0 (conservative). Only applied where the vendor publishes a cached rate.
agent_talk_share0.5Share of the call during which the agent is speaking. TTS is billed only on generated speech; STT, telephony and platform fees are billed on the whole call.
chars_per_audio_min1000Characters of text per minute of generated speech. Rime and ElevenLabs both state ~1,000 characters is about one minute of audio; Cartesia states 750-800. We use 1,000.

Speech-to-text

Streaming transcription, billed on the whole call.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
Sonioxstt-rt-v5 (real-time)$0.0020$2.00$0.12/hr (token-billed: $2.00/1M input audio tokens)
Diarization, language ID and formatting included per vendor.
verified2026-10-17 · source
AssemblyAIUniversal-Streaming (English)$0.0025$2.50$0.15/hr
Billed on session duration. Streaming diarization +$0.12/hr.
verified2026-10-17 · source
AssemblyAIUniversal-Streaming (Multilingual)$0.0025$2.50$0.15/hr
Billed on session duration.
verified2026-10-17 · source
InworldSTT 1 (On-Demand)$0.0025$2.50$0.15/hr on-demand; $0.10/hr on paid plansverified2026-10-17 · source
OpenAIgpt-4o-mini-transcribe (file/batch)$0.0030$3.00$0.003/min
Batch/file transcription; not a streaming model, so excluded from realtime 'cheapest' picks.
verified2026-10-17 · source
xAISpeech to Text (streaming)$0.0033$3.33$0.20/hr streaming; $0.10/hr RESTverified2026-10-17 · source
DeepgramNova-3 Monolingual, streaming$0.0048$4.80$0.0048/min
Promotional rate; regular $0.0077/min
verifiedpromo2026-10-17 · source
DeepgramNova-3 Multilingual, streaming$0.0058$5.80$0.0058/min
Promotional rate; regular $0.0092/min
verifiedpromo2026-10-17 · source
DeepgramFlux (English), streaming$0.0065$6.50$0.0065/min (Pay As You Go)
Billed per second. Growth plan rates are lower. Promotional rate; regular $0.0077/min
verifiedpromo2026-10-17 · source
ElevenLabsScribe v2 Realtime$0.0065$6.50$0.39/hr
Keyterms +$0.08/hr. Batch Scribe v2 is $0.22/hr.
verified2026-10-17 · source
CartesiaInk (streaming STT), Scale plan$0.0065$6.503 credits/sec; ~$0.39/hr on Scale plan (vendor figure)
Credit-based; effective rate depends on plan ($299/mo Scale). Lower plans cost more per credit.
verified2026-10-17 · source
AssemblyAIUniversal-3.6 Pro Realtime$0.0075$7.50$0.45/hr
Streaming billed on session duration, not speech duration.
verified2026-10-17 · source
DeepgramFlux Multilingual, streaming$0.0078$7.80$0.0078/minverified2026-10-17 · source
Google Gemini APIgemini-3.5-transcribe-live$0.0090$9.00~$0.009/min (vendor's blended estimate)
Token-billed; Google's page gives the per-minute estimate.
verified2026-10-17 · source
Google CloudSpeech-to-Text V2 Standard (0-500K min/mo)$0.0160$16.00$0.016/min
Tiered: $0.01 (500K-1M), $0.008 (1-2M), $0.004 (>2M min/mo).
verified2026-10-17 · source
OpenAIgpt-realtime-whisper (streaming)$0.0170$17.00$0.017/minverified2026-10-17 · source
OpenAIgpt-live-transcribe (streaming)$0.0170$17.00$0.017/minverified2026-10-17 · source

LLM

Per-token prices converted to a per-call-minute estimate using the assumptions below.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
OpenAIgpt-5-nano$0.0005$0.50$0.05 / $0.005 / $0.40 per 1Mverified2026-10-17 · source
Google Gemini APIgemini-2.5-flash-lite$0.0009$0.90$0.10 / $0.40 per 1Mverified2026-10-17 · source
OpenAIgpt-6-luna$0.0009$0.92$0.10 in / $0.01 cached / $0.50 out per 1M tokensverified2026-10-17 · source
OpenAIgpt-4o-mini$0.0013$1.34$0.15 / $0.075 / $0.60 per 1Mverified2026-10-17 · source
OpenAIgpt-5.6-luna$0.0019$1.89$0.20 / $0.02 / $1.20 per 1Mverified2026-10-17 · source
OpenAIgpt-5.4-nano$0.0019$1.90$0.20 / $0.02 / $1.25 per 1Mverified2026-10-17 · source
Google Gemini APIgemini-3.1-flash-lite$0.0024$2.36$0.25 / $1.50 per 1Mverified2026-10-17 · source
OpenAIgpt-5-mini$0.0025$2.48$0.25 / $0.025 / $2.00 per 1Mverified2026-10-17 · source
Google Gemini APIgemini-3.5-flash-lite$0.0030$3.00$0.30 / $2.50 per 1Mverified2026-10-17 · source
Google Gemini APIgemini-2.5-flash$0.0030$3.00$0.30 / $2.50 per 1Mverified2026-10-17 · source
OpenAIgpt-4.1-mini$0.0036$3.58$0.40 / $0.10 / $1.60 per 1Mverified2026-10-17 · source
Google Gemini APIgemini-3.8-flash$0.0069$6.90$0.75 in / $3.75 out per 1M
Rises to $1.50 / $7.50 per 1M on 2027-01-01 (per Google's page)
verifiedchanging2026-10-17 · source
OpenAIgpt-5.4-mini$0.0071$7.08$0.75 / $0.075 / $4.50 per 1Mverified2026-10-17 · source
AnthropicClaude Haiku 4.5 (legacy)$0.0092$9.20$1 in / $0.10 cache read / $5 out per 1Mverified2026-10-17 · source
xAIgrok-4.3 (<200k prompt)$0.0106$10.60$1.25 / $0.20 / $2.50 per 1Mverified2026-10-17 · source
xAIgrok-4.7 (<200k prompt)$0.0174$17.44$2 / $0.50 / $6 per 1Mverified2026-10-17 · source
OpenAIgpt-6-sol$0.0184$18.40$2 / $0.20 / $10 per 1Mverified2026-10-17 · source
AnthropicClaude Sonnet 5 (legacy)$0.0184$18.40$2 in / $0.20 cache read / $10 out per 1Mverified2026-10-17 · source
OpenAIgpt-5.6-terra$0.0189$18.88$2 / $0.20 / $12 per 1Mverified2026-10-17 · source
AnthropicClaude Sonnet 4.6 (legacy)$0.0276$27.60$3 in / $0.30 cache read / $15 out per 1Mverified2026-10-17 · source
AnthropicClaude Haiku 5.5n/an/aPage showed $0.50 and $2.50 per MTok without input/output captions
Unverified: labels missing in fetched page; Retell lists Claude 5.5 Haiku at $0.0032/min, which does not obviously reconcile. Re-check by hand.
unverified2026-10-17 · source
AnthropicClaude Sonnet 5.5n/an/aPage showed $2 and $10 per MTok without input/output captions
Unverified: labels missing in fetched page.
unverified2026-10-17 · source

Text-to-speech

Billed on generated speech only, so scaled by agent talk share.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
Google Gemini APIgemini-3.8-flash-lite-tts$0.0045$4.50$6 per 1M audio output tokens (~$0.0015 per 10s of audio)
Google's page says prices double on 2027-01-01. Derived: $0.0015 per 10s x 6 = $0.009 per minute of generated audio.
verifiedchangingderived2026-10-17 · source
SonioxText-to-Speech (real-time)$0.0058$5.83~$0.70 per hour of generated speech (vendor figure); $4/1M text tokens + $21.50/1M audio tokens
Derived: $0.70 / 60 = $0.0117 per minute of generated audio.
verifiedderived2026-10-17 · source
Google Gemini APIgemini-3.8-flash-tts$0.0067$6.75$9 per 1M audio output tokens (~$0.00225 per 10s of audio)
Google's page says prices double on 2027-01-01. Derived: $0.00225 per 10s x 6 = $0.0135 per minute of generated audio (25 audio tokens/sec).
verifiedchangingderived2026-10-17 · source
OpenAItts-1$0.0075$7.50$15 per 1M charactersverified2026-10-17 · source
DeepgramAura-1$0.0075$7.50$0.015 per 1K charactersverified2026-10-17 · source
InworldRealtime TTS-2 Flash (On-Demand)$0.0075$7.50$15 per 1M characters on-demand ($7-10 on paid plans)verified2026-10-17 · source
xAIText to Speech$0.0075$7.50$15 per 1M charactersverified2026-10-17 · source
InworldRealtime TTS-2 (On-Demand)$0.0125$12.50$25 per 1M characters on-demand (as low as $12.50 on Growth)verified2026-10-17 · source
OpenAItts-1-hd$0.0150$15.00$30 per 1M charactersverified2026-10-17 · source
DeepgramAura-2$0.0150$15.00$0.030 per 1K charactersverified2026-10-17 · source
RimeMist v3$0.0150$15.00$0.03 per 1K characters (~$0.03/min per vendor)verified2026-10-17 · source
CartesiaSonic (Startup plan credits)$0.0196$19.601 credit = 1 character; Startup $49/mo for 1.25M credits
Plan-based; you pay the subscription whether or not you use the credits. Derived: $49 / 1,250,000 credits = $0.0392 per 1K characters. Overage is $45/1M on Startup, $38/1M on Scale, $65/1M on Pro.
verifiedderived2026-10-17 · source
ElevenLabsFlash / Turbo$0.0200$20.00$0.04 per 1K charactersverified2026-10-17 · source
ElevenLabsEleven v4 Turbo$0.0200$20.00$0.04 per 1K charactersverified2026-10-17 · source
ElevenLabsEleven v3 Conversational$0.0200$20.00$0.04 per 1K charactersverified2026-10-17 · source
DeepgramFlux TTS$0.0225$22.50$0.045 per 1K charactersverified2026-10-17 · source
RimeCoda$0.0250$25.00$0.05 per 1K charactersverified2026-10-17 · source
ElevenLabsEleven v4$0.0400$40.00$0.08 per 1K characters (page: ~$0.08/min)verified2026-10-17 · source
ElevenLabsMultilingual v2$0.0400$40.00$0.08 per 1K charactersverified2026-10-17 · source
OpenAIgpt-4o-mini-ttsn/an/a$0.60/1M text input + $12/1M audio output tokens
Token price verified, but OpenAI's page does not state audio tokens per minute, so we do not publish a per-minute figure.
unverified2026-10-17 · source
PlayHTPlayHT APIn/an/aPricing page unreachable on 2026-10-17
Could not fetch; status of the product unknown.
unverified2026-10-17 · source

Speech-to-speech / realtime

Single models that hear and speak directly.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
Google Gemini APIgemini-3.8-live$0.0140$14.00Audio input $0.005/min, audio output $0.018/min; text $0.75 in / $4.50 out per 1M
Check Google's page: Gemini prices scheduled to change 2027-01-01. Per call-minute = $0.005 x 1 (caller audio, whole call) + $0.018 x agent talk share. Context/text tokens not included.
verifiedchangingderived2026-10-17 · source
OpenAIgpt-live-1$0.0500$50.00$0.05/min
Backend model and tools are billed separately (per OpenAI's page).
verified2026-10-17 · source
xAIgrok-voice-3.0$0.0800$80.00$0.08/min ($4.80/hr), $0.004 per text inputverified2026-10-17 · source
OpenAIgpt-realtime-2.1n/an/aAudio: $32 in / $0.40 cached / $64 out per 1M tokens
OpenAI does not publish audio tokens per minute, so no per-minute figure. For reference Retell resells it at $0.38/min.
verified2026-10-17 · source
OpenAIgpt-realtime-2.1-mini / gpt-realtime-minin/an/aAudio: $10 in / $0.30 cached / $20 out per 1M tokens
No published per-minute figure. Retell resells GPT Realtime mini at $0.07/min.
verified2026-10-17 · source

All-in platforms & orchestrators

Hosting or bundled per-minute fees. Read what each includes.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
VapiHosting$0.0500$50.00$0.05/min hosting, prepaid
STT/LLM/TTS providers passed through at cost (no markup per Vapi). Telephony transport billed by your telephony provider. Free tier: $5 credit, 4 concurrent calls, 1 number. Pro: 10% of hosting fee with $999/mo minimum.
verified2026-10-17 · source
Retell AIVoice Infra$0.0550$55.00$0.055/min voice infra + Retell-priced LLM, TTS, telephony
Retell prices components itself (see breakdown). Speech-to-text is not listed as a separate charge; we infer it is inside the $0.055 voice infra because Retell's published floor of $0.07/min equals $0.055 infra + $0.015 voice + the cheapest LLM. Knowledge base +$0.005/min, PII removal +$0.01/min. Numbers $2/mo; first 20 concurrent calls free, then $8 per concurrency/mo. Retell's own page range: $0.07-$0.31/min. Billed per second.
verified2026-10-17 · source
CartesiaManaged Agents$0.0600$60.00$0.06/min (+$0.014/min if using a Cartesia phone number)
Check Cartesia's page for what LLM usage is included.
verified2026-10-17 · source
TwilioConversationRelay$0.0700$70.00$0.07/min
You bring the LLM via websocket. Voice minutes billed on top.
verified2026-10-17 · source
DeepgramVoice Agent API (Standard)$0.0750$75.00$0.075/min; BYO LLM or BYO TTS $0.065; BYO LLM+TTS $0.050; Advanced $0.163
Telephony separate.
verified2026-10-17 · source
AssemblyAIVoice Agent API$0.0750$75.00$0.075/min, all-inclusive
Telephony separate.
verified2026-10-17 · source
ElevenLabsElevenLabs Agents$0.0800$80.00$0.08 per additional call minute (burst $0.16/min)
LLM billed separately; telephony billed by your own provider. Plans include minutes: Free 15, Starter $6/75, Creator $22/275, Pro $99/1,238, Scale $299/3,738, Business $990/12,375.
verified2026-10-17 · source
ElevenLabsSpeech Engine (STT+TTS for your agent)$0.0800$80.00$0.08/minverified2026-10-17 · source
BlandBland Build plan$0.1200$120.00$0.12/min + $299/mo, 50 concurrent calls
Transfer time $0.04/min. Telephony separate.
verified2026-10-17 · source
BlandBland Start plan$0.1400$140.00$0.14/min, $0 platform fee, 10 concurrent calls
Telephony separate (BYO or Bland's at pass-through). Transfer time $0.05/min.
verified2026-10-17 · source
SynthflowSynthflow Enterprisen/an/aContracts from $30,000/year; no public per-minute rateunverified2026-10-17 · source

Telephony (US)

Carrier minutes. Inbound and outbound differ.

VendorProduct$/call-min$/1,000 minVendor's priceStatusChecked
TelnyxElastic SIP, local inbound$0.0032$3.20from $0.0032/min; numbers $1.00/mo
'Starting at' rate. Voice API adds $0.002/min if you use it instead of raw SIP.
verified2026-10-17 · source
PlivoBrowser SDK / SIP calls$0.0033$3.30$0.0033/min each wayverified2026-10-17 · source
TwilioSIP interface / BYOC$0.0040$4.00$0.0040/min each wayverified2026-10-17 · source
TelnyxElastic SIP, local outbound$0.0050$5.00from $0.005/min
'Starting at' rate; varies by destination.
verified2026-10-17 · source
PlivoLocal, inbound$0.0055$5.50$0.0055/min; numbers $0.50/moverified2026-10-17 · source
TwilioLocal number, inbound$0.0085$8.50$0.0085/min + $1.15/mo numberverified2026-10-17 · source
PlivoLocal, outbound (US)$0.0115$11.50$0.0115/minverified2026-10-17 · source
TwilioUS/Canada, outbound$0.0140$14.00$0.0140/minverified2026-10-17 · source
CartesiaCartesia phone number$0.0140$14.00$0.014/min
Only with Cartesia Managed Agents.
verified2026-10-17 · source
TelnyxToll-free, inbound$0.0150$15.00from $0.015/minverified2026-10-17 · source
Retell AIRetell-provided Twilio number (US)$0.0150$15.00$0.015/min
Only when using Retell.
verified2026-10-17 · source
PlivoToll-free, inbound$0.0180$18.00$0.0180/minverified2026-10-17 · source
TwilioToll-free, inbound$0.0220$22.00$0.0220/min + $2.15/mo numberverified2026-10-17 · source