Voxtral pricing: what the Voxtral provider speech to text costs
A current rate for Voxtral could not be verified. Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself.

On this page
Voxtral pricing and what it covers
A current rate for Voxtral could not be verified. Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself.
Figures in the original research below are dated examples. Use the vendor’s current terms and the same workload when estimating your cost.
Voxtral is the Voxtral provider's speech to text, billed per minute through the the Voxtral provider API, with open weight models to self-host. This is how the options are priced, with figures flagged from launch coverage.
The current batch speech to text generation, the Voxtral provider's low cost transcription, billed per minute through the API.
The streaming option for live transcription, reported at roughly double the batch rate.
The July 2025 launch pricing, positioned at less than half the price of comparable APIs.
Voxtral Mini and Small open weight models to run yourself or through a third-party host, for compute cost only.
Speech models are billed per minute on the the Voxtral provider platform, alongside its language models.
the Voxtral provider's live pricing page links to the Voxtral docs rather than showing the rate inline, so these figures come from launch coverage. Verify the current per minute rate on the Voxtral provider.ai.
Estimate your total cost
For transcription workload, add fixed fees, metered usage and any separately priced services. Use the allowance in your plan and price any usage above it.
| Input | What to record |
|---|---|
| Workload | transcription workload |
| Billable items | Billed audio duration, live versus recorded mode, channels, redaction and other add-ons. |
| Terms | Currency, billing period, included usage, overages and contract minimum. |
| Pilot | Accuracy on names and numbers, final transcript timing and billed audio duration. |
Figures in the original research below are dated examples. Use the vendor’s current terms and the same workload when estimating your cost.
Voxtral is usage based, so the monthly cost is minutes of audio times the per minute rate, batch or realtime. Open weight models shift the cost to compute if you self-host.
What to test with Voxtral
Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself.
Transcribe the same calls with overlap, accents and names. On a live stream, track partial revisions and the final text separately.
The right Voxtral option depends on whether you transcribe recorded or live audio and whether you want the hosted API or to self-host an open model. Every option is low cost, so match it to the job.
At about $0.003 a minute, Transcribe 2 is the value pick for transcribing files and media in bulk, where the per-minute rate matters more than streaming latency and large archives need to stay cheap.
At about $0.006 a minute, the realtime model suits live captions and streaming transcription where the delay before words appear decides whether the experience feels usable, at double the batch rate.
Voxtral Mini and Small ship as open weights, so teams can run them on their own infrastructure for the cost of compute, which suits privacy-sensitive workloads and very high volume where a per-minute API would add up.
Teams already building on the Voxtral provider's language models that want speech billed on the same platform and the same invoice, rather than adding a separate speech vendor to the stack.
Voxtral keeps transcription low cost. The main choice is hosted API versus self-hosting an open model, and batch versus realtime for your latency needs.
Voxtral pricing calculator: estimate your monthly cost
Voxtral figure uses about $0.003 a minute for Transcribe 2 and $0.006 for Realtime, from launch coverage. Verify Voxtral rates on the Voxtral provider.ai. See current PyAI pricing
Fees and terms to confirm
- Ask which charges apply for billed audio duration, live versus recorded mode, channels, redaction and other add-ons.
- Confirm what is included in the base plan and what is metered separately.
- Ask how the vendor calculates minimums, overages and renewals.
- Rate not inline: the Voxtral provider's pricing page links to the Voxtral docs rather than showing the per minute rate, so confirm it before budgeting.
- Realtime costs more: streaming is reported at about double the batch rate, so live transcription is priced a step higher.
- Figures from coverage: the per minute numbers come from launch materials and news, so re-verify on the Voxtral provider.ai.
- Self-host compute: open weight models avoid API fees but you pay compute to run them.
- Focused on STT: Voxtral is speech to text, so text to speech and a full voice stack need other tools.
- Platform context: Voxtral sits within the Voxtral provider's broader platform, so language model usage is billed separately.
Voxtral is genuinely low cost for transcription. The main caution is that the live rate is not shown inline, so verify it on the Voxtral provider.ai.
Read customer evidence for your workload
Look for recent customer examples using the same product edition and deployment. Ask the vendor for a reference or a pilot if the published example leaves out your workflow.
We looked for real, attributed user reviews of Voxtral that mention pricing and the product. As of October 2026 we found no public user reviews on the sources below, so there is nothing to quote honestly. Voxtral is the Voxtral provider's open speech model rather than a standalone reviewed product, so the the Voxtral provider AI reviews on G2 are about the chat and coding models, not Voxtral. There is no public review base specific to the speech model to quote from.
No public user reviews found
Evaluate Voxtral on a short trial and a reference call rather than on ratings, since no verifiable user reviews exist yet.
Compare the same scope
Keep the workload, channel and required support consistent across options. A price for one component covers a different scope from a proposal for a complete deployment.
For PyAI’s billed products and telephony terms, see current PyAI pricing.
Both are low cost and bill per minute, so the difference is the stack. Voxtral is focused low cost speech to text from the Voxtral provider with open weight models. PyAI adds Speak for text to speech and Agents for phone calls under one PyAI account. The table compares Voxtral and PyAI Hear row by row for transcription buyers.
| Criteria | Voxtral | PyAI |
|---|---|---|
| Pricing model | the Voxtral provider API, billed per minute | See current PyAI pricing |
| Transcription rate | About $0.003 batch, $0.006 realtime | See current PyAI pricing |
| Monthly fee | $0, usage based | See current PyAI pricing |
| Free to start | Via the the Voxtral provider platform | See current PyAI pricing |
| Self-hostable weights | ✓ Voxtral Mini and Small to self-host | Hosted service |
| Text to speech from the same provider | No, STT only | See current PyAI pricing |
| Output for phone systems | Transcripts and JSON | ✓ PyAI renders WAV, PCM or G.711 audio for calls |
| Phone agent sold by the same provider | No | See current PyAI pricing |
| Pricing clarity | Rate not shown inline | ✓ Flat published per minute |
Choose Voxtral if you want
- Low-cost transcription from the Voxtral provider's hosted API, billed per minute of audio.
- Open-weight Voxtral models under Apache 2.0 that you can self-host on your own GPUs.
- Speech on the same platform and account as the Voxtral provider's language models.
Choose PyAI if you want
- Transcription at a flat $0.001/min plus a free credit
- Spoken audio in WAV, PCM or G.711, which a Voxtral STT call does not return
- Hear, Speak and Agents on one PyAI invoice, where Voxtral covers transcription only
In short, Voxtral is a strong pick when you want low cost transcription from the Voxtral provider and open weight models to self-host. PyAI fits when you want a flat published rate, a free credit to start, and text to speech and a phone agent on the same bill. Both keep transcription cheap, so the choice is about the wider stack and open models.
Check normal and peak usage
Estimate one normal billing period and one peak period. In each scenario record your billable units, included allowances, overages and any fixed fees. Compare the totals per completed task.
Figures in the original research below are dated examples. Use the vendor’s current terms and the same workload when estimating your cost.
Voxtral bills per minute with no plan, so cost scales with volume and whether you use batch or realtime. The table uses the Transcribe 2 rate against PyAI Hear at every volume.
| Use case | What you need | Voxtral option | Watch out for |
|---|---|---|---|
| Recorded files | Any volume | Transcribe 2 ($0.003) | Confirm the rate with the Voxtral provider |
| Live captions | Any volume | Realtime ($0.006) | Realtime costs more |
| Own infrastructure | Self-host | Mini or Small weights | Compute cost |
Voxtral vs PyAI monthly cost for every volume
Voxtral is shown at the Transcribe 2 rate of about $0.003 a minute. PyAI Hear prices the same minutes. Both are low cost and scale linearly, so the real difference is the free credit and the bundled text to speech. See current PyAI pricing
| Minutes a month | Voxtral option | Voxtral cost | PyAI Hear |
|---|---|---|---|
| 600 | Transcribe 2 | ~$1.80 | See current PyAI pricing |
| 3,000 | Transcribe 2 | ~$9 | See current PyAI pricing |
| 6,000 | Transcribe 2 | ~$18 | See current PyAI pricing |
| 12,000 | Transcribe 2 | ~$36 | See current PyAI pricing |
| 30,000 | Transcribe 2 | ~$90 | See current PyAI pricing |
Voxtral figures use the Transcribe 2 rate of about $0.003 a minute from launch coverage; realtime costs more. See current PyAI pricing
What to take into a vendor conversation
Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself.
Voxtral by the Voxtral provider is low cost speech to text, with Transcribe 2 from about $0.003 a minute, a realtime option around $0.006, and open weight models to self-host. The per minute rate is not shown inline on the Voxtral provider.ai, so these figures come from launch coverage. At 100 hours of audio, Voxtral is about $18. See current PyAI pricing
PyAI is not only a transcription model and does not offer open weights; it is voice infrastructure priced per minute, with speech to text, text to speech and a phone agent on one bill and a free credit to start.
Voxtral pricing FAQs
How much does Voxtral cost?
A current rate for Voxtral could not be verified. Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself.
What should I include in a Voxtral estimate?
Include billed audio duration, live versus recorded mode, channels, redaction and other add-ons. Use the same billing period across all quotes and ask how excess usage is charged.
How can I test Voxtral before committing?
Check the exact model, endpoint and distinction between hosted API billing and running an available model yourself. Transcribe the same calls with overlap, accents and names. On a live stream, track partial revisions and the final text separately.
How much does Voxtral cost per minute?
Voxtral, the Voxtral provider's speech to text, is low cost per minute. The current Transcribe 2 generation starts at about $0.003 a minute, with a realtime option reported at roughly double, around $0.006. The original Voxtral launched from about $0.001 a minute. Check the Voxtral provider's pricing page for the current Voxtral rate.
Confirm the current rate and what it includes with the vendor.
Does the Voxtral provider publish Voxtral pricing?
the Voxtral provider's pricing page states speech models are billed per minute but links out to the Voxtral docs rather than showing the rate inline, and the docs did not render a per minute number when checked. The figures here come from launch coverage and should be re-verified on the Voxtral provider.ai before budgeting.
Is Voxtral cheaper than Whisper?
the Voxtral provider positioned Voxtral at less than half the price of comparable APIs, and the hosted Whisper API has historically been about $0.006 a minute. Voxtral Transcribe 2 at about $0.003 a minute undercuts that, while open weight Voxtral models can be self-hosted for compute only.
Confirm the current rate and what it includes with the vendor.
Can I self-host Voxtral?
Yes. Voxtral ships open weight models, such as Voxtral Mini and Voxtral Small, that developers can self-host or run through third-party hosts. One host lists them on a per second basis, so self-hosting is an option for teams that want to run the models on their own infrastructure.
What is Voxtral Realtime?
Voxtral Realtime is the streaming option for live transcription, reported at about twice the batch rate, so roughly $0.006 a minute against $0.003 for Transcribe 2. As with the batch rate, confirm the current realtime price on the Voxtral provider's pricing page before relying on it.
Confirm the current rate and what it includes with the vendor.
What is an alternative to Voxtral with text to speech on one bill?
Voxtral is focused low cost speech to text from the Voxtral provider, so choose on whether you need the wider stack or just transcription. See current PyAI pricing
Related pricing and comparisons
Estimate PyAI phone costs
This calculator estimates PyAI charges for the phone setup you select. Use the pricing sections above for other providers. See the PyAI rate card for billing terms.
- AI usage · 150 min
- $7.50
- Managed telephony · 200 billable min
- $2.00
- Estimated monthly total
- $9.50
AI usage bills by the second. Managed telephony bills in whole-minute increments per call. This estimate assumes every call lasts the duration you enter.