Voice pricing, broken down. Know what you pay.
Start with 100 free test minutes. Production voice calls cost $0.055/min for the Ortavox platform, plus transcription, language model, voice and carrier usage. Includes 4 concurrent calls. All prices are in USD.
Free test
$0
100 test minutes
Build and test an agent in the WebRTC playground. No card required.
- 100 free test minutes
- Python & TypeScript SDK
- Browser WebRTC playground
- Call replay & transcripts
Pay-as-you-go
$0.055
per minute + provider pass-through
Production cloud. No monthly minimum, no seats, no lock-in.
- 4 concurrent calls included
- Additional concurrency quoted on request
- SIP via Twilio / Telnyx
- Bring your own model keys
- Turn-by-turn tracing & latency waterfall
- Sub-second median response
Enterprise
Custom
from $0.040/min at volume
Dedicated infrastructure or fully on-premise inside your network.
- Reserved compute, no shared queues
- On-premise on your GPUs (K8s / Docker)
- Zero audio egress
- Signed GDPR Article 28 DPA
- Named solutions engineer + SLA
Estimate your monthly voice costs
All amounts are in USD. Adjust provider assumptions to match your model and carrier. These are planning estimates, not a quote. The 100 free test minutes are separate from this production estimate.
Model estimates use the OrtaVox rate card below, updated October 12, 2026. STT uses audio minutes; LLM uses input, cached input and output tokens; TTS uses generated characters. Choose Custom / BYOK to enter your own average cost per call minute. Carrier defaults are illustrative; actual usage, channels and carrier destination can change the bill. For browser calls, set telephony to $0. BYOK changes who bills you; provider usage still costs money.
Estimated monthly total
$105.25
$0.1053 per call minute under these assumptions
- Ortavox platform · $0.055/min
- $55.00
- Transcription
- $5.00
- Language model
- $5.25
- Voice
- $30.00
- Telephony
- $10.00
- Optional usage services
- $0.00
- Fixed monthly costs
- $0.00
Includes 4 concurrent calls. Additional capacity requires a quote. Taxes are excluded. Number rental and optional services are included only when entered above. LLM and voice costs are converted into per-call-minute assumptions for this estimate. Providers may bill you separately from Ortavox.
Inference rate card
Model pricing, down to the unit
OrtaVox model pricing in USD. Model usage is separate from the $0.055/min platform fee and carrier costs. Each model has one published rate.
Updated October 12, 2026. BYOK follows your provider account. Cache-write charges and provider extras are separate. Available models depend on your agent configuration.
61 models · USD per 1 million tokens
| Model | Provider | Input | Cached input | Output |
|---|---|---|---|---|
| Claude Sonnet 4.6 | Anthropic | $3.000 | $0.300 | $15.000 |
| Claude Haiku 4.5 | Anthropic | $1.000 | $0.100 | $5.000 |
| Claude Opus 4.6 | Anthropic | $5.000 | $0.500 | $25.000 |
| Gemini 3 Flash | $0.500 | $0.050 | $3.000 | |
| Gemini 3.1 Flash Lite | $0.250 | $0.025 | $1.500 | |
| Gemini 3.1 Pro | $4.000 | $0.400 | $18.000 | |
| Gemini 3.5 Flash | $1.500 | $0.150 | $9.000 | |
| Gemini 3.5 Flash Lite | $0.300 | $0.030 | $2.500 | |
| Gemini 3.6 Flash | $1.500 | $0.150 | $7.500 | |
| Gemini 3.7 Flash | $0.750 | $0.075 | $3.750 | |
| Gemini 3.8 Flash | $0.750 | $0.075 | $3.750 | |
| Gemini 2.5 Flash | $0.300 | $0.030 | $2.500 | |
| Gemini 2.5 Flash-Lite | $0.100 | $0.010 | $0.400 | |
| Mistral Small 4 | Mistral | $0.150 | $0.015 | $0.600 |
| Mistral Medium 3.5 | Mistral | $1.500 | $0.150 | $7.500 |
| Mistral Large 3 | Mistral | $0.500 | $0.050 | $1.500 |
| ChatGPT Latest | OpenAI | $5.000 | $0.500 | $30.000 |
| GPT-4.1 | OpenAI | $2.000 | $0.500 | $8.000 |
| GPT-4.1 mini | OpenAI | $0.400 | $0.100 | $1.600 |
| GPT-4o | OpenAI | $2.500 | $1.250 | $10.000 |
| GPT-4o mini | OpenAI | $0.150 | $0.075 | $0.600 |
| GPT-5 | OpenAI | $1.250 | $0.125 | $10.000 |
| GPT-5 mini | OpenAI | $0.250 | $0.025 | $2.000 |
| GPT-5 nano | OpenAI | $0.050 | $0.005 | $0.400 |
| GPT-5.1 | OpenAI | $1.250 | $0.125 | $10.000 |
| GPT-5.2 | OpenAI | $1.750 | $0.175 | $14.000 |
| GPT-5.4 | OpenAI | $5.000 | $0.500 | $22.500 |
| GPT-5.4 mini | OpenAI | $0.750 | $0.075 | $4.500 |
| GPT-5.4 nano | OpenAI | $0.200 | $0.020 | $1.250 |
| GPT-5.5 | OpenAI | $10.000 | $1.000 | $45.000 |
| GPT-5.6 Luna | OpenAI | $0.200 | $0.020 | $1.200 |
| GPT-5.6 Sol | OpenAI | $4.000 | $0.400 | $20.000 |
| GPT-5.6 Terra | OpenAI | $2.000 | $0.200 | $12.000 |
| GPT-4.1 nano | OpenAI | $0.100 | $0.025 | $0.400 |
| ChatGPT Latest | Azure | $5.000 | $0.500 | $30.000 |
| GPT-4.1 | Azure | $2.000 | $0.500 | $8.000 |
| GPT-4.1 mini | Azure | $0.400 | $0.100 | $1.600 |
| GPT-4.1 nano | Azure | $0.100 | $0.030 | $0.400 |
| GPT-4o | Azure | $2.500 | $1.250 | $10.000 |
| GPT-4o mini | Azure | $0.150 | $0.075 | $0.600 |
| GPT-5 | Azure | $1.250 | $0.130 | $10.000 |
| GPT-5 mini | Azure | $0.250 | $0.030 | $2.000 |
| GPT-5 nano | Azure | $0.050 | $0.010 | $0.400 |
| GPT-5.1 | Azure | $1.250 | $0.130 | $10.000 |
| GPT-5.2 | Azure | $1.750 | $0.180 | $14.000 |
| GPT-5.4 | Azure | $5.000 | $0.500 | $22.500 |
| GPT-5.5 | Azure | $10.000 | $1.000 | $45.000 |
| GPT-5.6 Luna | Azure | $0.200 | $0.020 | $1.200 |
| GPT-5.6 Sol | Azure | $4.000 | $0.400 | $20.000 |
| GPT-5.6 Terra | Azure | $2.000 | $0.200 | $12.000 |
| DeepSeek-V4.1 Flash | Baseten | $0.300 | $0.030 | $1.200 |
| Gemma 4 31B | LiveKit | $0.400 | $0.200 | $1.200 |
| GPT OSS 120B | Baseten | $0.100 | N/A | $0.500 |
| GPT OSS 120B | Groq | $0.150 | $0.075 | $0.600 |
| Grok 4.20 | SpaceXAI | $2.000 | N/A | $6.000 |
| Grok 4.20 Reasoning | SpaceXAI | $2.000 | N/A | $6.000 |
| Grok 4.20 Multi-Agent | SpaceXAI | $2.000 | N/A | $6.000 |
| Grok 4.3 | SpaceXAI | $1.250 | $0.200 | $2.500 |
| Grok 4.5 | SpaceXAI | $2.000 | $0.500 | $6.000 |
| Grok 4.6 | SpaceXAI | $2.000 | $0.500 | $6.000 |
| Grok 4.7 | SpaceXAI | $2.000 | $0.500 | $6.000 |
Detailed component pricing
Every part of a voice call
Your total combines the Ortavox platform fee with actual model and carrier usage. Provider costs are passed through at cost. With your own keys, those providers bill you directly.
Total call cost = platform + transcription + language model + voice + telephony
Provider units differ. Per-call-minute estimates convert their usage into an average; they are not fixed model prices.
| Component | Price / billing unit | What it covers | Billed by |
|---|---|---|---|
| Ortavox voice platform | $0.055 / connected minute | Orchestration, WebRTC/SIP media, turn-taking, tool calls, tracing and transcripts. | Ortavox |
| Speech recognition (STT) | Selected provider’s audio rate | Transcribes incoming audio. Model, language, channels and transcription options affect cost. | At cost or your provider account |
| Language model (LLM) | Selected provider’s token rate | Input and output tokens. Longer prompts, conversation history and tool responses can increase usage. | At cost or your provider account |
| Speech synthesis (TTS) | Selected provider’s character or audio rate | Generates the agent’s voice. Model, voice and amount of generated speech affect cost. | At cost or your provider account |
| Phone carrier | Carrier rate by destination | Inbound or outbound phone transport through your SIP carrier. Browser calls do not use a phone carrier. | Your carrier |
Choose the deployment that fits
| What’s included | Free test | Pay-as-you-go | Enterprise |
|---|---|---|---|
| Platform price | $0 for testing | $0.055/min + providers | Custom; volume rates from $0.040/min |
| Included test usage | 100 test minutes | Production usage billed separately | Agreed during onboarding |
| Concurrent production calls | Test environment | 4 included | Reserved capacity by agreement |
| Deployment | Browser playground | Managed cloud | Dedicated cloud or on-premise |
| Provider billing | Test offer | At cost or BYOK | Provider accounts or local models |
| Phone connectivity | Browser testing | Twilio / Telnyx SIP | Carrier and topology by agreement |
| SDKs | Python & TypeScript | Python & TypeScript | Python & TypeScript |
| Support & service terms | Community | Standard support | Named engineer and agreed SLA |
Capacity and additional costs
Four active calls can run at the same time across your agents. Each call contributes usage separately. Extra concurrency prices and enterprise services are quoted for your workload.
Concurrent calls
4 included
More simultaneous calls: request a capacity quote.
Phone numbers
Carrier pricing
Number rental is separate from per-minute carrier usage.
Provider extras
Provider pricing
Optional transcription features and other provider services follow their own rate cards.
Dedicated infrastructure
Custom quote
Compute, region and reserved capacity are specified in your agreement.
On-premise deployment
Annual licence quote
Your infrastructure and GPU costs are separate from the Ortavox licence.
Enterprise support & SLA
Custom agreement
Response times, availability commitments and onboarding are agreed for your deployment.
How usage adds up
One 10-minute call
$0.55 platform fee
10 × $0.055. Transcription, language model, voice and any carrier usage are additional.
Four calls for 10 minutes each
$2.20 platform fee
4 × 10 = 40 billed call minutes, even when calls overlap. This uses all 4 included concurrent slots.
1,000 production minutes
$55 platform fee
1,000 × $0.055. The estimator adds your provider assumptions and any monthly costs you enter.
Pricing questions
- What exactly does the per-minute rate cover?
- Orchestration, WebRTC and SIP media, turn-taking, tool calling, tracing and storage of transcripts. Model and telephony providers are passed through at cost, or billed directly by those providers when you bring your own keys. BYOK does not make model usage free.
- How many calls can run at the same time?
- The pay-as-you-go offer includes 4 concurrent calls: up to 4 active calls at once, across your agents. This is capacity, not 4 free calls or minutes. Each connected call contributes its own usage. Contact us for additional capacity and pricing.
- Is there a monthly minimum or seat fee?
- No. Pay-as-you-go is metered per second of connected call time and billed monthly. Seats in the console are unlimited.
- How does volume pricing work?
- Above roughly 50K minutes per month rates step down toward $0.040/min. Reserved capacity contracts are quoted per workload.
- Does the calculator show my Ortavox invoice or total spend?
- It shows estimated total spend, including providers. With BYOK, your Ortavox invoice covers the platform and agreed Ortavox services; providers and carriers send separate bills. Add monthly number rental and quoted services to the estimate to budget for both.
- Are phone numbers and taxes included?
- The platform rate excludes phone-number rental, provider extras and taxes. Your carrier sets number rental and call rates. Enter any known monthly costs in the calculator; taxes remain excluded.
- Do browser calls need a telephony budget?
- Browser WebRTC calls do not use a phone carrier. Set the carrier rate to zero for those calls. The Ortavox platform and model usage still apply.
- Can I run Ortavox entirely on my own infrastructure?
- Yes. On-premise deployments run in your Kubernetes or Docker environment on your GPUs, licensed annually. No audio leaves your network.