A Recurring Recording Library
At 21,600 base minutes, Speech is Cheap Subscription costs $20.00. AssemblyAI Universal-2 costs $54.00, or 2.7× our bill. Deepgram Nova-3 Monolingual PAYG costs $92.88, or 4.64× our bill.
See the Base TotalsSpeech-to-Text API Pricing Comparison
Compare ten providers on the bill for your recorded audio. Start with monthly minutes, then include word timestamps, speaker diarization and audio labels. See which plan fits your workload and how live transcription pricing differs from processing existing recordings.
Recorded Audio
Compare the Monthly Bill
Four workloads, with each product and plan shown separately. Bold totals mark the lowest listed cost for that workload and feature set.
| Provider and Product | 600 min 10 hours | 6000 min 100 hours | 21,600 min 360 hours | 60,000 min 1000 hours |
|---|---|---|---|---|
| Speech is Cheap PAYG | $1.20 | $12.00 | $43.20 | $120.00 |
| Speech is Cheap Subscription | $20.00 | $20.00 | $20.00 | $55.56 |
| AssemblyAI Universal-2 | $1.50 | $15.00 | $54.00 | $150.00 |
| AssemblyAI Universal-3.5 Pro | $2.10 | $21.00 | $75.60 | $210.00 |
| Deepgram Nova-3 Monolingual PAYG | $2.58 | $25.80 | $92.88 | $258.00 |
| Deepgram Nova-3 Multilingual PAYG | $3.12 | $31.20 | $112.32 | $312.00 |
| OpenAI whisper-1 (Sunset) | $3.60 | $36.00 | $129.60 | $360.00 |
| OpenAI GPT-Transcribe | $2.70 | $27.00 | $97.20 | $270.00 |
| OpenAI 4o Mini (Estimate, Sunset) | $1.80 | $18.00 | $64.80 | $180.00 |
| OpenAI 4o Diarize (Estimate, Sunset) | $3.60 | $36.00 | $129.60 | $360.00 |
| Rev AI Reverb English | $2.00 | $20.00 | $72.00 | $200.00 |
| Rev AI Reverb Other Languages | $3.00 | $30.00 | $108.00 | $300.00 |
| Google Cloud Standard, Chirp 3 | $9.60 | $96.00 | $345.60 | $960.00 |
| Google Cloud Dynamic Batch, Chirp 3 | $1.80 | $18.00 | $64.80 | $180.00 |
| Amazon Transcribe Standard Batch, US East | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech Batch | $1.80 | $18.00 | $64.80 | $180.00 |
| Azure Speech Fast | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech MAI-2 Preview Promo Promo Ends 2026-12-31 | $1.00 | $10.00 | $36.00 | $100.00 |
| Gladia Recorded Starter | $6.10 | $61.00 | $219.60 | $610.00 |
| Speechmatics Melia 1, Early Access | $1.29 | $12.90 | $46.44 | $116.10 |
| Speechmatics Standard | $2.40 | $24.00 | $86.40 | $216.00 |
| Speechmatics Enhanced | $4.00 | $40.00 | $144.00 | $360.00 |
Transcription and Word Timestamps
| Provider and Product | 600 min 10 hours | 6000 min 100 hours | 21,600 min 360 hours | 60,000 min 1000 hours |
|---|---|---|---|---|
| Speech is Cheap PAYG | $1.80 | $18.00 | $64.80 | $180.00 |
| Speech is Cheap Subscription | $20.30 | $23.00 | $30.80 | $85.56 |
| AssemblyAI Universal-2 | $1.50 | $15.00 | $54.00 | $150.00 |
| AssemblyAI Universal-3.5 Pro | $2.10 | $21.00 | $75.60 | $210.00 |
| Deepgram Nova-3 Monolingual PAYG | $2.58 | $25.80 | $92.88 | $258.00 |
| Deepgram Nova-3 Multilingual PAYG | $3.12 | $31.20 | $112.32 | $312.00 |
| OpenAI whisper-1 (Sunset) | $3.60 | $36.00 | $129.60 | $360.00 |
| OpenAI GPT-Transcribe | -- | -- | -- | -- |
| OpenAI 4o Mini (Estimate, Sunset) | -- | -- | -- | -- |
| OpenAI 4o Diarize (Estimate, Sunset) | -- | -- | -- | -- |
| Rev AI Reverb English | $2.00 | $20.00 | $72.00 | $200.00 |
| Rev AI Reverb Other Languages | $3.00 | $30.00 | $108.00 | $300.00 |
| Google Cloud Standard, Chirp 3 | -- | -- | -- | -- |
| Google Cloud Dynamic Batch, Chirp 3 | -- | -- | -- | -- |
| Amazon Transcribe Standard Batch, US East | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech Batch | $1.80 | $18.00 | $64.80 | $180.00 |
| Azure Speech Fast | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech MAI-2 Preview Promo Promo Ends 2026-12-31 | $1.00 | $10.00 | $36.00 | $100.00 |
| Gladia Recorded Starter | $6.10 | $61.00 | $219.60 | $610.00 |
| Speechmatics Melia 1, Early Access | $1.29 | $12.90 | $46.44 | $116.10 |
| Speechmatics Standard | $2.40 | $24.00 | $86.40 | $216.00 |
| Speechmatics Enhanced | $4.00 | $40.00 | $144.00 | $360.00 |
Transcription and Speaker Diarization
| Provider and Product | 600 min 10 hours | 6000 min 100 hours | 21,600 min 360 hours | 60,000 min 1000 hours |
|---|---|---|---|---|
| Speech is Cheap PAYG | $2.40 | $24.00 | $86.40 | $240.00 |
| Speech is Cheap Subscription | $20.60 | $26.00 | $41.60 | $115.56 |
| AssemblyAI Universal-2 | $1.70 | $17.00 | $61.20 | $170.00 |
| AssemblyAI Universal-3.5 Pro | $2.30 | $23.00 | $82.80 | $230.00 |
| Deepgram Nova-3 Monolingual PAYG | $2.58 | $25.80 | $92.88 | $258.00 |
| Deepgram Nova-3 Multilingual PAYG | $3.12 | $31.20 | $112.32 | $312.00 |
| OpenAI whisper-1 (Sunset) | -- | -- | -- | -- |
| OpenAI GPT-Transcribe | -- | -- | -- | -- |
| OpenAI 4o Mini (Estimate, Sunset) | -- | -- | -- | -- |
| OpenAI 4o Diarize (Estimate, Sunset) | $3.60 | $36.00 | $129.60 | $360.00 |
| Rev AI Reverb English | $2.00 | $20.00 | $72.00 | $200.00 |
| Rev AI Reverb Other Languages | $3.00 | $30.00 | $108.00 | $300.00 |
| Google Cloud Standard, Chirp 3 | $9.60 | $96.00 | $345.60 | $960.00 |
| Google Cloud Dynamic Batch, Chirp 3 | $1.80 | $18.00 | $64.80 | $180.00 |
| Amazon Transcribe Standard Batch, US East | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech Batch | $1.80 | $18.00 | $64.80 | $180.00 |
| Azure Speech Fast | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech MAI-2 Preview Promo Promo Ends 2026-12-31 | $1.00 | $10.00 | $36.00 | $100.00 |
| Gladia Recorded Starter | $6.10 | $61.00 | $219.60 | $610.00 |
| Speechmatics Melia 1, Early Access | $1.29 | $12.90 | $46.44 | $116.10 |
| Speechmatics Standard | $2.40 | $24.00 | $86.40 | $216.00 |
| Speechmatics Enhanced | $4.00 | $40.00 | $144.00 | $360.00 |
Transcription, Words and Speakers
| Provider and Product | 600 min 10 hours | 6000 min 100 hours | 21,600 min 360 hours | 60,000 min 1000 hours |
|---|---|---|---|---|
| Speech is Cheap PAYG | $3.00 | $30.00 | $108.00 | $300.00 |
| Speech is Cheap Subscription | $20.90 | $29.00 | $52.40 | $145.56 |
| AssemblyAI Universal-2 | $1.70 | $17.00 | $61.20 | $170.00 |
| AssemblyAI Universal-3.5 Pro | $2.30 | $23.00 | $82.80 | $230.00 |
| Deepgram Nova-3 Monolingual PAYG | $2.58 | $25.80 | $92.88 | $258.00 |
| Deepgram Nova-3 Multilingual PAYG | $3.12 | $31.20 | $112.32 | $312.00 |
| OpenAI whisper-1 (Sunset) | -- | -- | -- | -- |
| OpenAI GPT-Transcribe | -- | -- | -- | -- |
| OpenAI 4o Mini (Estimate, Sunset) | -- | -- | -- | -- |
| OpenAI 4o Diarize (Estimate, Sunset) | -- | -- | -- | -- |
| Rev AI Reverb English | $2.00 | $20.00 | $72.00 | $200.00 |
| Rev AI Reverb Other Languages | $3.00 | $30.00 | $108.00 | $300.00 |
| Google Cloud Standard, Chirp 3 | -- | -- | -- | -- |
| Google Cloud Dynamic Batch, Chirp 3 | -- | -- | -- | -- |
| Amazon Transcribe Standard Batch, US East | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech Batch | $1.80 | $18.00 | $64.80 | $180.00 |
| Azure Speech Fast | $3.60 | $36.00 | $129.60 | $360.00 |
| Azure Speech MAI-2 Preview Promo Promo Ends 2026-12-31 | $1.00 | $10.00 | $36.00 | $100.00 |
| Gladia Recorded Starter | $6.10 | $61.00 | $219.60 | $610.00 |
| Speechmatics Melia 1, Early Access | $1.29 | $12.90 | $46.44 | $116.10 |
| Speechmatics Standard | $2.40 | $24.00 | $86.40 | $216.00 |
| Speechmatics Enhanced | $4.00 | $40.00 | $144.00 | $360.00 |
Transcription and Audio Labels
| Provider and Product | 600 min 10 hours | 6000 min 100 hours | 21,600 min 360 hours | 60,000 min 1000 hours |
|---|---|---|---|---|
| Speech is Cheap PAYG | $1.32 | $13.20 | $47.52 | $132.00 |
| Speech is Cheap Subscription | $20.06 | $20.60 | $22.16 | $61.56 |
| AssemblyAI Universal-2 | -- | -- | -- | -- |
| AssemblyAI Universal-3.5 Pro | -- | -- | -- | -- |
| Deepgram Nova-3 Monolingual PAYG | -- | -- | -- | -- |
| Deepgram Nova-3 Multilingual PAYG | -- | -- | -- | -- |
| OpenAI whisper-1 (Sunset) | -- | -- | -- | -- |
| OpenAI GPT-Transcribe | -- | -- | -- | -- |
| OpenAI 4o Mini (Estimate, Sunset) | -- | -- | -- | -- |
| OpenAI 4o Diarize (Estimate, Sunset) | -- | -- | -- | -- |
| Rev AI Reverb English | -- | -- | -- | -- |
| Rev AI Reverb Other Languages | -- | -- | -- | -- |
| Google Cloud Standard, Chirp 3 | -- | -- | -- | -- |
| Google Cloud Dynamic Batch, Chirp 3 | -- | -- | -- | -- |
| Amazon Transcribe Standard Batch, US East | -- | -- | -- | -- |
| Azure Speech Batch | -- | -- | -- | -- |
| Azure Speech Fast | -- | -- | -- | -- |
| Azure Speech MAI-2 Preview Promo Promo Ends 2026-12-31 | -- | -- | -- | -- |
| Gladia Recorded Starter | -- | -- | -- | -- |
| Speechmatics Melia 1, Early Access | -- | -- | -- | -- |
| Speechmatics Standard | $2.40 | $24.00 | $86.40 | $216.00 |
| Speechmatics Enhanced | $4.00 | $40.00 | $144.00 | $360.00 |
-- means no verified price for the selected product or feature combination. It never means free. This applies to every table and the linked calculators.
Separate the Base Rate from Add-ons
Words are word-level timestamps. Speakers are voice-based diarization. Included features add no charge to the selected product's base rate.
Rates are displayed to six decimal places. Totals use the unrounded source rates and applicable volume tiers. Billing rules and output differences are explained below.
What Changes the Lowest Bill?
Named workloads make the price differences reproducible. Each example points back to the product rows above.
Conversations with Words and Speakers
At 6000 minutes with both features, Speech is Cheap Subscription costs $29.00 and PAYG costs $30.00. AssemblyAI Universal-2 costs $17.00; Speechmatics Melia 1 costs $12.90. Included features change the result.
See the Add-on TablesA Small Base-Only Workload
At 600 base minutes, Speech is Cheap PAYG costs $1.20. Azure MAI-2 Preview Promo costs $1.00. That promotional price ends 2026-12-31 and has preview restrictions. A $20.00 subscription is unnecessary at this volume.
Review Azure's Preview LimitsFile duration matters too. For 100 six-second clips, Speech is Cheap PAYG bills 100 minutes and costs $0.20. AssemblyAI Universal-2 prorates the ten minutes of audio and costs $0.03. A whole-minute workload cannot predict that short-clip bill.
Choose a live service for microphone input. Compare deployment location, required contracts, language support and the results on your own recordings before selecting a provider. Speech is Cheap's recorded API does not replace live input or a self-hosted deployment.
Live Audio
Price Streaming as a Separate Service
These reference rates apply to live input. They do not enter the recorded-workload totals or the cost multiples above.
| Provider and Product | Base / Minute | Billing and Product Boundary |
|---|---|---|
| Speech is Cheap Recorded Audio Only | -- | No live-input product. Recorded-job progress SSE does not accept a live microphone. Speech is Cheap Job Requests |
| AssemblyAI Universal-Streaming | $0.0025 $0.15 / hour | English or multilingual. Billed on WebSocket-open time, including silence. Diarization adds $0.002 per minute. AssemblyAI Live Pricing ; AssemblyAI Session Billing |
| AssemblyAI Universal-3.5 Pro Realtime | $0.0075 $0.45 / hour | Billed on WebSocket-open time. Diarization adds $0.002 per minute. AssemblyAI Live Pricing ; AssemblyAI Session Billing |
| Deepgram Nova-3 Monolingual PAYG | $0.0048 $0.288 / hour | Promotional rate; no end date published. Regular rate $0.0077 per minute. Live add-ons are separate. Deepgram Streaming Pricing |
| Deepgram Nova-3 Multilingual PAYG | $0.0058 $0.348 / hour | Promotional rate; no end date published. Regular rate $0.0092 per minute. Live add-ons are separate. Deepgram Streaming Pricing |
| OpenAI GPT-Live-Transcribe | $0.017 $1.02 / hour | Live transcription billed by audio duration. Voice-agent generation and live translation are different products. OpenAI GPT-Live-Transcribe Pricing |
| Rev AI Streaming API | -- | Streaming is documented, but a distinct current streaming rate was not established from the public pricing page. Confirm it with Rev AI. Rev AI Pricing ; Rev AI Frequently Asked Questions |
| Google Cloud V2 Standard Recognition | $0.016 $0.96 / hour | First 500,000 minutes; marginal volume tiers apply. Check the streaming model and language before selecting features. Google Cloud V2 Pricing ; Google Cloud Streaming Recognition |
| Amazon Transcribe Standard Streaming, US East | $0.010002 $0.60012 / hour | Northern Virginia. Exact retail rate is $0.0001667 per second, with no minimum. Call Analytics is separate. Amazon Transcribe Streaming Price List |
| Azure Speech S1 Standard Realtime, East US | $0.016667 $1 / hour | Per-second billing. Diarization adds $0.005 per minute. Custom models and commitment tiers are separate. Azure East US Live Retail Prices ; Azure Speech Billing and Features |
| Gladia Live Starter | $0.0125 $0.75 / hour | Public Starter rate. Growth requires an upfront commitment; enterprise terms are quoted separately. Gladia Live Starter Pricing |
| Speechmatics Realtime Standard | $0.004 $0.24 / hour | 20% marginal discount above 30,000 minutes per model and processing type. Melia 1 is batch-only. Speechmatics Realtime Pricing ; Speechmatics Model Comparison |
| Speechmatics Realtime Enhanced | $0.007167 $0.43 / hour | 20% marginal discount above 30,000 minutes per model and processing type. The batch Enhanced rate differs. Speechmatics Realtime Pricing |
Calculate Your Provider Comparison
Use each provider's calculator for other volumes and supported add-on combinations. Its alternative page also covers capabilities and integration work.
AssemblyAIDeepgramOpenAI Whisper APIRev AIGoogle Cloud Speech-to-TextAmazon TranscribeAzure SpeechGladiaSpeechmatics
See the 360-Hour Audio Bill Infographic for a shareable example of selected base-transcription prices.
Sources and Verification
Speech is Cheap publishes this comparison and sells one of the products shown. Recorded prices retain their 2026-09-14 and 2026-09-15 checks; new live-price and credit sources were checked on . Review rates monthly and before publication or a material update. A check date records the evidence review, not a guarantee that the offer remains available. Price comparisons do not establish equivalent accuracy, latency, hosting or compliance.
- Amazon Transcribe Free Tier Terms
Checked
- Amazon Transcribe Input and Output
Checked
- Amazon Transcribe Pricing
Checked
- Amazon Transcribe Streaming Price List
Checked
- Amazon Transcribe US East Price List
Checked
- AssemblyAI Billing and Pricing
Checked
- AssemblyAI Live Pricing
Checked
- AssemblyAI Pricing
Checked
- AssemblyAI Session Billing
Checked
- Azure Batch Transcription Overview
Checked
- Azure East US Live Retail Prices
Checked
- Azure Fast Transcription
Checked
- Azure MAI-Transcribe Features and Preview Limits
Checked
- Azure Speech Billing and Features
Checked
- Azure Speech East US Retail Prices
Checked
- Azure Speech Pricing
Checked
- Deepgram Pricing
Checked
- Deepgram Recorded Audio API
Checked
- Deepgram Streaming Pricing
Checked
- Gladia Live Starter Pricing
Checked
- Gladia Pricing
Checked
- Gladia Transcription Pricing Explained
Checked
- Google Cloud Batch Transcription
Checked
- Google Cloud BatchRecognize API
Checked
- Google Cloud Chirp 3 Features and Limits
Checked
- Google Cloud Recognition Configuration
Checked
- Google Cloud Speech Languages and Models
Checked
- Google Cloud Speech-to-Text Pricing
Checked
- Google Cloud Streaming Recognition
Checked
- Google Cloud V2 Pricing
Checked
- Microsoft MAI-Transcribe-2 Launch and Promotional Price
Checked
- OpenAI API Pricing
Checked
- OpenAI Diarization Model
Checked
- OpenAI File Transcription Guide
Checked
- OpenAI GPT-Live-Transcribe Pricing
Checked
- OpenAI GPT-Transcribe Model
Checked
- OpenAI Model Deprecations
Checked
- OpenAI Open-Source Whisper
Checked
- OpenAI Whisper Model and Price
Checked
- Rev AI Changelog
Checked
- Rev AI Frequently Asked Questions
Checked
- Rev AI Pricing
Checked
- Rev AI Pricing
Checked
- Rev AI Transcript Features
Checked
- Speech is Cheap Job Requests
Checked
- Speech is Cheap Pricing
Checked
- Speechmatics Model Comparison
Checked
- Speechmatics Pricing and Volume Discounts
Checked
- Speechmatics Realtime Pricing
Checked
Choose a Plan for Your Recorded Audio
Review Speech is Cheap's included minutes, overage and add-ons, then sign up for Subscription or PAYG.
Billing Units and Assumptions
All prices are in USD; promotional credits retain their original currencies below. Recorded examples use single-channel, ten-minute files in one billing month: 60, 600, 2160 or 6000 files, totaling 600, 6000, 21,600 or 60,000 minutes. Selected add-ons apply to every file. Use a supported language and model; the Google examples use supported English audio. These totals exclude taxes, storage, transfer, trial credits, negotiated contracts and optional features that are not selected.
Divide a vendor's hourly price by 60 to get its minute rate. Multiply a per-second price by 60. For each file, apply that provider's duration rounding and minimum first, then sum the billable minutes. Charge each marginal volume band at its own rate and add any fixed plan fee and selected add-ons. The displayed unit rates are rounded; the totals are calculated with full precision and rounded to cents at the end.
Speech is Cheap rounds each file up to a whole minute for base transcription and all enabled add-ons. PAYG total = billed minutes × (base rate + selected add-on rates). Subscription total = $20.00 + max(0, billed minutes − 21,600) × $0.000926 + billed minutes × selected subscription add-on rates. Unused included minutes expire at the end of the billing period. Add-ons are charged even within the base allowance. Speech is Cheap Pricing
A cost multiple = the named competitor's total ÷ the named Speech is Cheap plan's total for the same workload and features. For the library example, $54.00 ÷ $20.00 = 2.7×. Percent saved = 100 × (competitor total − Speech is Cheap total) ÷ competitor total. These calculations describe the named scenarios, not a universal savings range or equal accuracy.
Trial credits can reduce a first bill, and commitments can change recurring rates. They are separate from these standard paid-usage totals. Speech is Cheap's $0 upfront PAYG signup is not a recurring free API allowance. Check each provider's eligibility, expiration, data-use terms and minimum spend before applying a credit or discount to your budget.
Credits and Commitments
- AssemblyAI
- The pricing page advertises up to 11,100 recorded minutes or 19,980 streaming minutes for evaluation. Check the selected product and account terms; these are not assumed to reset monthly. AssemblyAI Live Pricing
- Deepgram
- A $200 introductory credit is separate from paid usage. Growth starts at $4000 per year in prepaid credits. Deepgram Streaming Pricing
- Rev AI
- Trial credits equal 300 minutes of Reverb ASR and can be used across products. Enterprise rates depend on a volume agreement. Rev AI Pricing
- Google Cloud
- The 60 free monthly minutes listed for V1 are not a V2 allowance. New-customer cloud credits and commitment discounts are separate offers. Google Cloud V2 Pricing
- Amazon Transcribe
- The pricing page lists 60 free minutes per month for the first 12 months, subject to eligibility. Unused minutes do not roll over. Amazon Transcribe Free Tier Terms
- Azure Speech
- F0 lists 300 free live-transcription minutes per month. This is not a free batch or fast-transcription allowance. Paid commitment tiers have separate minimum spends. Azure Speech Billing and Features
- Gladia
- Starter advertises a one-time €50 credit with no monthly reset. It is not converted to dollars or deducted from these totals. Growth requires a separate upfront commitment. Gladia Live Starter Pricing
- Speechmatics
- New accounts start with $100 in credit. Data-use opt-in discounts and enterprise agreements are separate from the automatic marginal volume discount included here. Speechmatics Realtime Pricing
AssemblyAI
Recorded audio is prorated to the second. Standard paid usage draws from a prepaid balance. Hourly recorded rates are divided by 60. Word timestamps are included; speaker diarization is a separate charge. Live sessions use connection time instead of the recorded-file duration. AssemblyAI Pricing ; AssemblyAI Billing and Pricing
Deepgram
Deepgram Nova-3 PAYG costs $0.0043 per minute for monolingual audio and $0.0052 for multilingual audio, with word timing and speaker diarization included. Billing is per second. Growth pricing requires an upfront commitment and is excluded. Deepgram Pricing ; Deepgram Recorded Audio API
OpenAI
Hosted Whisper costs $0.006 per minute and GPT-Transcribe costs $0.0045. GPT-4o Mini and GPT-4o Diarize columns use minute-based estimates of $0.003 and $0.006 for token-priced models; actual bills can vary. These GPT-4o models and whisper-1 are scheduled to sunset on February 26, 2027. OpenAI Whisper Model and Price ; OpenAI API Pricing ; OpenAI GPT-Transcribe Model ; OpenAI Diarization Model ; OpenAI Model Deprecations
The listed OpenAI models do not have a verified single-call configuration that returns both word timestamps and speaker annotations. Separate inference passes, external diarization, translation and self-hosting are outside these estimates. OpenAI File Transcription Guide ; OpenAI Open-Source Whisper
Rev AI
Rev AI lists Reverb English at $0.20 per audio hour and other supported languages at $0.30, equivalent to approximately $0.003333 and $0.005 per minute. Calculations divide the exact hourly rates by 60. Each job is rounded up to the next second with a 15-second minimum. Standard word timing and diarization are included; premium diarization and human transcription are excluded. Rev AI Pricing ; Rev AI Transcript Features ; Rev AI Changelog
Google Cloud
Google V2 standard tiers are $0.016 per minute for the first 500,000 minutes, $0.01 for the next 500,000, $0.008 for the next 1,000,000, and $0.004 thereafter. Each rate applies only to usage in its tier. Dynamic batch costs $0.003 per minute and can defer processing for up to 24 hours. Billing rounds each request to one-second increments. Google Cloud Speech-to-Text Pricing ; Google Cloud BatchRecognize API
These Chirp 3 estimates use a supported English locale with speaker diarization included. Word-timing combinations remain unpriced because the model page describes optional timestamps under an unsupported-features heading. The generic API configuration field does not resolve that conflict. Google Cloud Chirp 3 Features and Limits ; Google Cloud Recognition Configuration ; Google Cloud Speech Languages and Models
Chirp 3 batch guidance recommends recordings of 1–60 minutes, or up to 20 minutes with word timing. These model-specific limits take precedence over the generic 480-minute batch ceiling. Cloud Storage request charges are additional. Google Cloud Batch Transcription ; Google Cloud Chirp 3 Features and Limits
Amazon Transcribe
AWS standard batch in US East, Northern Virginia, costs $0.0001 per second, equivalent to $0.006 per minute, with one-second billing and no job minimum. The public price list published 2026-09-11 has a single base tier. Standard word timing and speaker diarization are included. Amazon Transcribe Pricing ; Amazon Transcribe US East Price List ; Amazon Transcribe Input and Output
Channel separation has no extra charge in either service. Speech is Cheap assigns one speaker per channel with is_split_channel; AWS standard pricing covers up to two channels. This is separate from identifying voices with Parse Speakers. S3 requests and additional AWS services are outside these totals. Amazon Transcribe Pricing ; Amazon Transcribe Input and Output ; Speech is Cheap Job Requests ; Speech is Cheap Pricing
Azure Speech
Azure East US lists batch at $0.18 per hour and fast transcription at $0.36, equivalent to $0.003 and $0.006 per minute. MAI-2 preview pricing is $0.10 per hour through 2026-12-31. Its listed standard rate is $0.006 per minute, 3.6× the promotion. At 100,000 minutes that is $600 instead of $166.67. Use the listed standard rate to budget for 2027, subject to future Microsoft pricing changes. Azure Speech East US Retail Prices ; Azure Speech Pricing ; Microsoft MAI-Transcribe-2 Launch and Promotional Price
Word timing and speaker diarization are included in the selected paths. MAI-2 enhanced mode documents diarization failures at about 15 minutes and longer, so its combined-feature estimates apply only to shorter recordings. Batch is deferred processing; Fast and MAI-2 use the fast transcription path. Check API-version and preview availability requirements before deployment. Azure MAI-Transcribe Features and Preview Limits ; Azure Batch Transcription Overview ; Azure Fast Transcription ; Azure Speech Pricing
Gladia
Gladia recorded Starter lists $0.61 per hour, approximately $0.010167 per minute, with word timing and speaker diarization included. Calculations divide the exact hourly rate by 60. Growth advertises prices as low as $0.20 per hour with an upfront commitment; the required commitment is unpublished, so Growth is excluded. Gladia Pricing ; Gladia Transcription Pricing Explained
Speechmatics
Speechmatics lists batch rates of $0.129, $0.24 and $0.40 per hour for Melia 1, Standard and Enhanced. Calculations divide those exact rates by 60. The automatic 20% discount applies only to usage above 500 hours, or 30,000 minutes, per model and processing type each month. The separate 33% model-training discount requires data-use opt-in and is excluded. Speechmatics Pricing and Volume Discounts
All three models include word timing and speaker diarization. Melia 1 is early access and has a reduced feature set. The table shows words, speakers and audio labels separately; combine them in the calculator. Speechmatics Model Comparison ; Speechmatics Pricing and Volume Discounts
Audio-label totals compare Speech is Cheap segment labels with Speechmatics Standard and Enhanced Audio Events. Their event taxonomies and output structures differ; Melia 1 is unpriced for this selection. Channel separation is different from voice-based diarization. OpenAI's selected products cannot be combined into one verified words-and-speakers price, and Google's conflicting word-timing documentation leaves those combinations unpriced.