Speech-to-text API pricing compared
Transcribing an hour of recorded audio costs from $0.04 with Groq Whisper Large V3 Turbo to $0.61 with Gladia Starter (async), taking each of 12 vendors' cheapest option at list prices: $0.0007 to $0.0102 a minute. Live (streaming) transcription runs from $0.15 an hour with AssemblyAI Universal-Streaming to $1.02 with OpenAI gpt-live-transcribe across the 10 vendors that price it; Groq and Rev AI list no streaming price.
Regular pay-as-you-go prices from each vendor's own pricing page, converted to dollars per hour of audio. Amazon's prices are for US East (N. Virginia) and Azure's for East US. Price only: accuracy, languages and latency differ, and we don't rate them.
Transcription cost calculator
Enter your monthly hours of audio and whether it's recorded files or a live stream. Every option is re-ranked for that volume.
Cheapest for 100 hours of batch audio a month: Groq Whisper Large V3 Turbo at $4 ($0.04 an hour).
| Option | Per month | Per hour | Per minute |
|---|---|---|---|
| Groq Whisper Large V3 Turbo | $4 | $0.04 | $0.0007 |
| Groq Whisper Large V3 | $11.10 | $0.111 | $0.0019 |
| Speechmatics Melia 1 | $12 | $0.12 | $0.002 |
| AssemblyAI Universal-2 | $15 | $0.15 | $0.0025 |
| OpenAI gpt-4o-mini-transcribe Token-billed; per-minute figure is an estimate. | $18 | $0.18 | $0.003 |
| Google Cloud V2 dynamic batch | $18 | $0.18 | $0.003 |
| Azure AI Speech Batch transcription | $18 | $0.18 | $0.003 |
| Rev AI Reverb (English) | $20 | $0.20 | $0.0033 |
| AssemblyAI Universal-3.5 Pro | $21 | $0.21 | $0.0035 |
| ElevenLabs Scribe v2 First 270 minutes a month free. | $21.01 | $0.21 | $0.0035 |
| Speechmatics Standard | $23 | $0.23 | $0.0038 |
| Deepgram Nova-3 Monolingual | $25.80 | $0.258 | $0.0043 |
| OpenAI gpt-transcribe | $27 | $0.27 | $0.0045 |
| Deepgram Whisper Large | $28.80 | $0.288 | $0.0048 |
| AssemblyAI Sync API | $30 | $0.30 | $0.005 |
| Rev AI Reverb (other languages) | $30 | $0.30 | $0.005 |
| Rev AI Whisper Large (English) | $30 | $0.30 | $0.005 |
| Gemini API Gemini 3.5 Transcribe Token-billed; per-minute figure is an estimate. | $30.60 | $0.306 | $0.0051 |
| Deepgram Nova-3 Multilingual | $31.20 | $0.312 | $0.0052 |
| OpenAI Whisper (whisper-1) | $36 | $0.36 | $0.006 |
| OpenAI gpt-4o-transcribe Token-billed; per-minute figure is an estimate. | $36 | $0.36 | $0.006 |
| OpenAI gpt-4o-transcribe-diarize Token-billed; per-minute figure is an estimate. | $36 | $0.36 | $0.006 |
| Amazon Transcribe Standard | $36 | $0.36 | $0.006 |
| Azure AI Speech Fast transcription | $36 | $0.36 | $0.006 |
| Azure AI Speech MAI Transcribe-1.5 | $36 | $0.36 | $0.006 |
| Azure AI Speech LLM Speech | $36 | $0.36 | $0.006 |
| Speechmatics Enhanced | $38 | $0.38 | $0.0063 |
| Gladia Starter (async) | $61 | $0.61 | $0.0102 |
| Google Cloud V1 with data logging First 60 minutes a month free. | $95.04 | $0.95 | $0.0158 |
| Google Cloud V2 standard | $96 | $0.96 | $0.016 |
| Google Cloud V1 without data logging First 60 minutes a month free. | $143 | $1.43 | $0.0238 |
51 pay-as-you-go options from 12 vendors at regular list prices, with monthly volume tiers and free monthly minutes that don't expire. Promotions, one-off credits, commitments and specialist models (medical, custom, human) are left out of the ranking.
Every vendor at 10, 100, 1,000 and 10,000 hours a month
Each vendor's cheapest option at each volume, as cost per audio hour, sorted at 1,000 hours a month. Free monthly minutes count; one-off credits don't.
Recorded audio (batch)
| Vendor | 10 hours | 100 hours | 1,000 hours | 10,000 hours |
|---|---|---|---|---|
| Groq Whisper Large V3 Turbo | $0.04 $0.0007/min | $0.04 $0.0007/min | $0.04 $0.0007/min | $0.04 $0.0007/min |
| Speechmatics Melia 1 | $0.12 $0.002/min | $0.12 $0.002/min | $0.12 $0.002/min | $0.12 $0.002/min |
| AssemblyAI Universal-2 | $0.15 $0.0025/min | $0.15 $0.0025/min | $0.15 $0.0025/min | $0.15 $0.0025/min |
| OpenAI gpt-4o-mini-transcribe | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min |
| Google Cloud V2 dynamic batch | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min |
| Azure AI Speech Batch transcription | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min | $0.18 $0.003/min |
| Rev AI Reverb (English) | $0.20 $0.0033/min | $0.20 $0.0033/min | $0.20 $0.0033/min | $0.20 $0.0033/min |
| ElevenLabs Scribe v2 | $0.121 $0.002/min | $0.21 $0.0035/min | $0.219 $0.0037/min | $0.22 $0.0037/min |
| Deepgram Nova-3 Monolingual | $0.258 $0.0043/min | $0.258 $0.0043/min | $0.258 $0.0043/min | $0.258 $0.0043/min |
| Gemini API Gemini 3.5 Transcribe | $0.306 $0.0051/min | $0.306 $0.0051/min | $0.306 $0.0051/min | $0.306 $0.0051/min |
| Amazon Transcribe Standard | $0.36 $0.006/min | $0.36 $0.006/min | $0.36 $0.006/min | $0.36 $0.006/min |
| Gladia Starter (async) | $0.61 $0.0102/min | $0.61 $0.0102/min | $0.61 $0.0102/min | $0.61 $0.0102/min |
Live audio (streaming)
| Vendor | 10 hours | 100 hours | 1,000 hours | 10,000 hours |
|---|---|---|---|---|
| AssemblyAI Universal-Streaming | $0.15 $0.0025/min | $0.15 $0.0025/min | $0.15 $0.0025/min | $0.15 $0.0025/min |
| Speechmatics Linden 1 (voice agents) | $0.20 $0.0033/min | $0.20 $0.0033/min | $0.20 $0.0033/min | $0.20 $0.0033/min |
| ElevenLabs Scribe v2 Realtime | $0.293 $0.0049/min | $0.38 $0.0063/min | $0.389 $0.0065/min | $0.39 $0.0065/min |
| Deepgram Nova-3 Monolingual | $0.462 $0.0077/min | $0.462 $0.0077/min | $0.462 $0.0077/min | $0.462 $0.0077/min |
| Gemini API Gemini 3.5 Transcribe Live | $0.536 $0.0089/min | $0.536 $0.0089/min | $0.535 $0.0089/min | $0.535 $0.0089/min |
| Amazon Transcribe Standard | $0.60 $0.01/min | $0.60 $0.01/min | $0.60 $0.01/min | $0.60 $0.01/min |
| Gladia Starter (real-time) | $0.75 $0.0125/min | $0.75 $0.0125/min | $0.75 $0.0125/min | $0.75 $0.0125/min |
| Google Cloud V1 with data logging | $0.864 $0.0144/min | $0.95 $0.0158/min | $0.959 $0.016/min | $0.90 $0.015/min, V2 standard |
| Azure AI Speech Real-time transcription | $1 $0.0167/min | $1 $0.0167/min | $1 $0.0167/min | $1 $0.0167/min |
| OpenAI gpt-live-transcribe | $1.02 $0.017/min | $1.02 $0.017/min | $1.02 $0.017/min | $1.02 $0.017/min |
Groq and Rev AI publish no streaming price, so they're missing from the live table. Google's V2 price falls after 500,000 minutes (about 8,333 hours) a month; every other price here is flat.
Every model, cheapest first
| Vendor and model | Audio | Per minute | Per hour |
|---|---|---|---|
| Groq Whisper Large V3 Turbo Transcription only, no translation. | batch | $0.0007 | $0.04 |
| Groq Whisper Large V3 | batch | $0.0019 | $0.111 |
| Speechmatics Melia 1 Multilingual; the real-time version is marked coming soon. | batch | $0.002 | $0.12 |
| AssemblyAI Universal-2 | batch | $0.0025 | $0.15 |
| AssemblyAI Universal-Streaming English only. Billed on connection time, not audio sent. | streaming | $0.0025 | $0.15 |
| AssemblyAI Universal-Streaming Multilingual Billed on connection time, not audio sent. | streaming | $0.0025 | $0.15 |
| Azure AI Speech Batch transcription | batch | $0.003 | $0.18 |
| Google Cloud V2 dynamic batch Processed at lower urgency in exchange for the discount. | batch | $0.003 | $0.18 |
| OpenAI gpt-4o-mini-transcribe Estimate: token-billed. | batch | $0.003 | $0.18 |
| Rev AI Reverb (English) | batch | $0.0033 | $0.20 |
| Speechmatics Linden 1 (voice agents) Shown as $0.20, struck through, and $0.15, with no end date. | streaming | $0.0033 | $0.20 |
| AssemblyAI Universal-3.5 Pro | batch | $0.0035 | $0.21 |
| ElevenLabs Scribe v2 | batch | $0.0037 | $0.22 |
| Speechmatics Standard | batch | $0.0038 | $0.23 |
| Speechmatics Standard | streaming | $0.0038 | $0.23 |
| Deepgram Nova-3 Monolingual | batch | $0.0043 | $0.258 |
| OpenAI gpt-transcribe | batch | $0.0045 | $0.27 |
| Deepgram Whisper Large | batch | $0.0048 | $0.288 |
| AssemblyAI Sync API Universal-3.5 Pro in one call, for clips up to 2 minutes. | batch | $0.005 | $0.30 |
| Rev AI Reverb (other languages) | batch | $0.005 | $0.30 |
| Rev AI Whisper Large (English) | batch | $0.005 | $0.30 |
| Gemini API Gemini 3.5 Transcribe Estimate: token-billed. | batch | $0.0051 | $0.306 |
| Deepgram Nova-3 Multilingual | batch | $0.0052 | $0.312 |
| Amazon Transcribe Standard | batch | $0.006 | $0.36 |
| Azure AI Speech Fast transcription Synchronous transcription of files. | batch | $0.006 | $0.36 |
| Azure AI Speech MAI Transcribe-1.5 Microsoft's page doesn't say whether MAI Transcribe takes files or live audio; its price list bills MAI Transcribe-2's promotion on a Fast Transcription meter, so we group both with file transcription. | batch | $0.006 | $0.36 |
| Azure AI Speech LLM Speech Shares Fast Transcription's price and billing code. | batch | $0.006 | $0.36 |
| OpenAI Whisper (whisper-1) | batch | $0.006 | $0.36 |
| OpenAI gpt-4o-transcribe Estimate: token-billed. | batch | $0.006 | $0.36 |
| OpenAI gpt-4o-transcribe-diarize Estimate: token-billed. | batch | $0.006 | $0.36 |
| Speechmatics Enhanced | batch | $0.0063 | $0.38 |
| ElevenLabs Scribe v2 Realtime | streaming | $0.0065 | $0.39 |
| Speechmatics Enhanced | streaming | $0.0067 | $0.40 |
| AssemblyAI Universal-3.6 Pro Realtime Billed on connection time, not audio sent. | streaming | $0.0075 | $0.45 |
| Deepgram Nova-3 Monolingual | streaming | $0.0077 | $0.462 |
| Deepgram Flux English Built for voice agents, with turn detection. | streaming | $0.0077 | $0.462 |
| Deepgram Flux Multilingual Built for voice agents, with turn detection. | streaming | $0.0078 | $0.468 |
| Gemini API Gemini 3.5 Transcribe Live Estimate: token-billed. | streaming | $0.0089 | $0.536 |
| Deepgram Nova-3 Multilingual | streaming | $0.0092 | $0.552 |
| Amazon Transcribe Standard | streaming | $0.01 | $0.60 |
| Gladia Starter (async) | batch | $0.0102 | $0.61 |
| Gladia Starter (real-time) | streaming | $0.0125 | $0.75 |
| Google Cloud V2 standard Volume tiers apply to each month's minutes per account. | batch | $0.016 | $0.96 |
| Google Cloud V2 standard Volume tiers apply to each month's minutes per account. | streaming | $0.016 | $0.96 |
| Google Cloud V1 with data logging First 60 minutes a month free. Data logging lets Google use your audio. | batch | $0.016 | $0.96 |
| Google Cloud V1 with data logging First 60 minutes a month free. Data logging lets Google use your audio. | streaming | $0.016 | $0.96 |
| Azure AI Speech Real-time transcription | streaming | $0.0167 | $1 |
| OpenAI gpt-live-transcribe | streaming | $0.017 | $1.02 |
| OpenAI gpt-realtime-whisper | streaming | $0.017 | $1.02 |
| Google Cloud V1 without data logging First 60 minutes a month free. | batch | $0.024 | $1.44 |
| Google Cloud V1 without data logging First 60 minutes a month free. | streaming | $0.024 | $1.44 |
List price for the first volume tier. Google's V1 options include 60 free minutes a month and ElevenLabs' include 4.5 hours (Scribe v2) or 2.5 hours (Realtime); the calculator and the tables above count those.
Free tiers and credits
| Vendor | What's free | How often |
|---|---|---|
| OpenAI | No free allowance or credit is listed on OpenAI's pricing page. | None |
| Google Cloud | The V1 API's first 60 minutes a month are free; the V2 table lists no free minutes. | Every month |
| Google Cloud | New Google Cloud customers get $300 of free credit to try any Google Cloud product. | Once, at sign-up |
| Gemini API | The free tier charges nothing for input or output, but Google may use free-tier content to improve its products, which it doesn't on the paid tier. | Free plan |
| Amazon Transcribe | New accounts get 60 minutes a month free, for the first 12 months only. | Monthly, first 12 months |
| Azure AI Speech | A separate free tier (F0) gives 5 audio hours a month of real-time transcription; batch isn't supported on it. | Separate free tier |
| Deepgram | Every new account gets $200 of free credit, which doesn't expire until used. | Once, at sign-up |
| AssemblyAI | Sign-up comes with $50 of free credit and no card required; AssemblyAI says that covers up to 185 hours of pre-recorded or 333 hours of streaming transcription. | Once, at sign-up |
| ElevenLabs | The Free / pay-as-you-go column of the API price table includes 4 hours 30 minutes of Scribe v2 and 2 hours 30 minutes of Scribe v2 Realtime. | Every month |
| Groq | The free plan transcribes up to 7,200 audio seconds an hour and 28,800 a day (8 hours) per Whisper model at no charge, within its rate limits. | Free plan |
| Rev AI | New accounts get free credits worth 5 hours of Reverb transcription, usable on any product. | Once, at sign-up |
| Speechmatics | New accounts get $100 of free credit, no card required. | Once, at sign-up |
| Gladia | New accounts get 50€ of free credit once, with no monthly reset; Gladia puts it at roughly 80+ hours of pre-recorded or 60+ hours of real-time transcription. | Once, at sign-up |
One-off credits are the bigger prize for a trial; monthly allowances matter for small, steady use. See the cheapest speech-to-text API for how far each credit goes.
How each vendor counts the seconds
Per-second billing only matters for short clips. For 10,000 clips of 5 seconds (13.9 hours of audio), each vendor's cheapest batch option costs:
| Vendor | Billing rule | Billed per clip | 10,000 clips |
|---|---|---|---|
| Groq Whisper Large V3 Turbo | Every request is billed for at least 10 seconds. | 10 s | $1.11 |
| Speechmatics Melia 1 | Usage is charged down to the second at the hourly rate. | 5 s | $1.67 |
| Google Cloud V2 dynamic batch | Billed per second: each request is rounded up to the nearest second. | 5 s | $2.50 |
| Azure AI Speech Batch transcription | Billed per second of audio sent. | 5 s | $2.50 |
| Deepgram Nova-3 Monolingual | Billed per second, with no rounding up. | 5 s | $3.58 |
| Amazon Transcribe Standard | Billed per second with no minimum. | 5 s | $5 |
| Rev AI Reverb (English) | Rounded up to the nearest second, with a 15-second minimum per file. | 15 s | $8.33 |
| OpenAI gpt-4o-mini-transcribe | OpenAI's pricing page doesn't say how partial minutes are counted. | Not stated | Can't tell |
| Gemini API Gemini 3.5 Transcribe | Billed in tokens: Google counts 25 audio tokens per second of input audio, plus the text tokens returned. | Not stated | Can't tell |
| AssemblyAI Universal-2 | Pre-recorded audio is billed per hour of audio submitted; streaming is billed for the time the connection is open, idle time included. AssemblyAI doesn't state a rounding increment. | Not stated | Can't tell |
| ElevenLabs Scribe v2 | ElevenLabs says speech to text is billed per audio minute; it doesn't say how partial minutes are counted. | Not stated | Can't tell |
| Gladia Starter (async) | Gladia's pricing page doesn't state a billing increment; it runs on a prepaid credit wallet. | Not stated | Can't tell |
Speaker labels (diarization)
Telling speakers apart is free on some APIs and an add-on on others:
| Vendor | Recorded audio | Live audio |
|---|---|---|
| OpenAI | Separate model | Not listed |
| Google Cloud | Not listed | Not listed |
| Gemini API | Included | Not listed |
| Amazon Transcribe | Included | Included |
| Azure AI Speech | Included | +$0.30 an hour |
| Deepgram | Included | +$0.12 an hour |
| AssemblyAI | +$0.02 an hour | +$0.12 an hour |
| ElevenLabs | Not listed | Not listed |
| Groq | Not listed | No streaming price |
| Rev AI | Not listed | No streaming price |
| Speechmatics | Included | Included |
| Gladia | Included | Included |
"Not listed" means the pricing page doesn't price or mention it; the feature may still exist. OpenAI labels speakers with a separate model, gpt-4o-transcribe-diarize.
Promotions running now
We rank regular prices. These are lower for now:
- Azure AI Speech MAI Transcribe-2 (batch): $0.0017 a minute ($0.10 an hour), the only price published, until 31 December 2026.
- Deepgram Nova-3 Monolingual (streaming): $0.0048 a minute ($0.288 an hour), against a regular $0.0077, with no end date given.
- Deepgram Flux English (streaming): $0.0065 a minute ($0.39 an hour), against a regular $0.0077, with no end date given.
- Deepgram Nova-3 Multilingual (streaming): $0.0058 a minute ($0.348 an hour), against a regular $0.0092, with no end date given.
- Speechmatics Linden 1 (voice agents) (streaming): $0.0025 a minute ($0.15 an hour), against a regular $0.0033, with no end date given.
Token-billed models
Some models bill tokens rather than minutes, so the per-minute price is an estimate:
- OpenAI gpt-4o-mini-transcribe: Billed in tokens ($1.25 per million input tokens, $5.00 per million output tokens); $0.003 a minute is OpenAI's own estimate.
- OpenAI gpt-4o-transcribe: Billed in tokens ($2.50 per million input tokens, $10.00 per million output tokens); $0.006 a minute is OpenAI's own estimate.
- OpenAI gpt-4o-transcribe-diarize: Billed in tokens like gpt-4o-transcribe; $0.006 a minute is OpenAI's own estimate. Labels speakers.
- Gemini API Gemini 3.5 Transcribe: Billed in tokens: $2.00 per million audio input tokens and $12.00 per million output tokens. Google's own basis (25 audio tokens a second, 175 output tokens a minute) gives $0.0051 a minute; Google rounds it to about $0.005.
- Gemini API Gemini 3.5 Transcribe Live: Billed in tokens: $3.50 per million audio input tokens and $21.00 per million output tokens. On Google's basis that is $0.0089 a minute; Google rounds it to about $0.009.
Priced but not ranked
Specialist products, kept out of the comparison because they do a different job:
- Google Cloud Medical dictation and conversation (batch): $0.078 a minute ($4.68 an hour). For medical audio.
- Google Cloud Medical dictation and conversation (streaming): $0.078 a minute ($4.68 an hour). For medical audio.
- Amazon Transcribe Medical (batch): $0.075 a minute ($4.50 an hour). For medical audio.
- Amazon Transcribe Medical (streaming): $0.075 a minute ($4.50 an hour). For medical audio.
- Azure AI Speech Custom speech (batch): $0.0038 a minute ($0.225 an hour). Plus $0.0538 an hour per hosted model and $10 per compute hour of training.
- Azure AI Speech Custom speech (real-time): $0.02 a minute ($1.20 an hour). Plus $0.0538 an hour per hosted model and $10 per compute hour of training.
- AssemblyAI Dictation API (batch): $0.0103 a minute ($0.62 an hour). Returns cleaned-up text from clips of up to 120 seconds.
- ElevenLabs Scribe v2 Medical (batch): $0.0037 a minute ($0.22 an hour). For medical audio.
- Rev AI Human transcription (batch): $1.99 a minute ($119 an hour). Transcribed by people.
- Speechmatics Oak 1 Medical (batch): $0.0025 a minute ($0.15 an hour). For medical audio.
Head to head
- Deepgram vs AssemblyAI
- Deepgram vs Whisper API
- AssemblyAI vs Whisper API
- Deepgram vs ElevenLabs Scribe
- AssemblyAI vs ElevenLabs Scribe
- AWS Transcribe vs Whisper API
- Google Speech-to-Text vs Whisper API
- Azure Speech to Text vs Whisper API
- AssemblyAI vs Speechmatics
- Speechmatics vs Deepgram
- Gladia vs Deepgram
- Groq Whisper vs OpenAI Whisper
- AWS Transcribe vs Google Speech-to-Text
- ElevenLabs Scribe vs Whisper API
- AWS Transcribe vs Deepgram
- Deepgram vs Azure Speech to Text
Sources
Every vendor publishes pay-as-you-go prices; none blocked us. Pages read on 6 October 2026, each with its quoted text on the vendor's page here:
- OpenAI: developers.openai.com/api/docs/pricing.md, developers.openai.com/api/docs/pricing, developers.openai.com/api/docs/models/whisper-1.md, developers.openai.com/api/docs/models/gpt-transcribe.md, developers.openai.com/api/docs/models/gpt-live-transcribe.md, developers.openai.com/api/docs/guides/speech-to-text.md
- Google Cloud: cloud.google.com/speech-to-text/pricing, cloud.google.com/free
- Gemini API: ai.google.dev/gemini-api/docs/pricing
- Amazon Transcribe: aws.amazon.com/transcribe/pricing/, b0.p.awsstatic.com/pricing/2.0/meteredUnitMaps/transcribe/USD/current/transcribe.json
- Azure AI Speech: azure.microsoft.com/en-us/pricing/details/speech/, prices.azure.com/api/retail/prices
- Deepgram: deepgram.com/pricing
- AssemblyAI: www.assemblyai.com/pricing, www.assemblyai.com/pricing.md
- ElevenLabs: elevenlabs.io/pricing/api
- Groq: console.groq.com/docs/speech-to-text, console.groq.com/docs/models, console.groq.com/docs/rate-limits
- Rev AI: www.rev.ai/pricing
- Speechmatics: www.speechmatics.com/pricing
- Gladia: www.gladia.io/pricing
Text to speech instead? See text-to-speech API pricing. How we convert and rank: methodology.
Prices on this page were checked against the official pricing pages on .