Methodology
Every figure on this site is computed from one data file in the site’s repository. Prices were last updated on 2026-10-23.
Speech to text
Your monthly minutes are converted to the provider’s billing unit (minutes or hours). With whole-volume tiers, the tier your total falls in prices every unit. With graduated tiers, each band of usage is priced at its own rate. Optional features such as speaker labels are added when you switch them on and the provider charges for them.
Text to speech
Your monthly characters are converted to thousands or millions of characters. Providers with per-character pricing are tiered the same way as speech to text. For plan-based pricing we pick the cheapest plan that includes your whole volume, or the largest plan plus its overage rate if none does.
Voice agents
Per-minute services are multiplied by your minutes. When your concurrent calls exceed the provider’s limit and a burst rate is published, the excess share of minutes is billed at the burst rate. Token-priced services are converted to minutes using the factors below, so those rows are estimates and are labelled as such.
What is not included
- Free-tier credits and promotional credits.
- Taxes, support plans, enterprise discounts and committed-spend contracts.
- Language model charges for voice agents, unless the provider bundles them into the per-minute price.
Sources
| Provider | Pricing page | Scenes | Read on |
|---|---|---|---|
| ElevenLabs | elevenlabs.io | Speech to text, Text to speech, Voice agents | 2026-10-23 |
| Deepgram | deepgram.com | Speech to text, Text to speech, Voice agents | 2026-10-23 |
| Mistral Voxtral | docs.mistral.ai | Speech to text, Text to speech, Voice agents | 2026-10-23 |
| OpenAI | developers.openai.com | Speech to text, Text to speech, Voice agents | 2026-10-23 |
| AssemblyAI | www.assemblyai.com | Speech to text, Voice agents | 2026-10-23 |
| Google Cloud | cloud.google.com more | Speech to text, Text to speech | 2026-10-23 |
| Azure | azure.microsoft.com | Speech to text, Text to speech | 2026-10-23 |
Token conversion factors
| Model | Input tokens per minute | Output tokens per minute |
|---|---|---|
| OpenAI GPT-Realtime-2.1 | 600 | 1200 |
| OpenAI GPT-Realtime-2.1 Mini | 600 | 1200 |
Currency
All sources publish prices in US dollars, so no conversion is applied.
Ranking and affiliate links
Providers are ordered by estimated monthly cost only. Some “Visit” links are affiliate links and may earn a commission. The ranking code never receives the link, so it cannot affect the order.
Disclaimer
Prices change and public pages can be incomplete. Treat the figures as estimates and confirm with the provider before you commit. Corrections are listed on the price changes page.