How does ElevenLabs API pricing work?
ElevenLabs API requests deduct usage from your account, but API rates are model- and product-specific rather than identical to website generation. As of 2026-08-09, the public API rate card lists text-to-speech at $0.05 per 1,000 characters for Flash/Turbo and $0.10 per 1,000 for Multilingual v2/v3; other products use their own units. Plan and concurrency limits still matter.
Why — the first-principles explanation
ElevenLabs API pricing has two layers that should not be collapsed. Operationally, an API request made with your key deducts credits from your account. Economically, the public API pricing page expresses current usage in dollars by product and model. Website generation therefore is not a safe shortcut for estimating an API workload.
For text-to-speech, the current public rate card lists $0.05 per 1,000 input characters for Flash/Turbo and $0.10 per 1,000 for Multilingual v2/v3. Speech-to-text, music, voice processing, sound effects and dubbing use different units such as audio hours, minutes or generations. Duration is only an estimate derived from a script and speaking pace; there is no universal credits-per-minute API rule.
Plan level affects the amount of included usage and concurrency, the number of requests that can run at once. Model choice, input length and product determine usage cost, while concurrency determines whether the service can handle a traffic spike. Check both the live API rate card and the current model-limit table before committing to a production budget. These rates were verified on 2026-08-09 and can change.
An example that makes it click
Think of the API like a utility bill with several meters. Text-to-speech counts input characters, transcription counts audio time, and sound effects count generations. Flash/Turbo and Multilingual v2/v3 are also different tariffs on the text meter. Your account balance pays the bill, while your plan's concurrency limit controls how many machines can run at the same time. A single credits-to-minutes formula would ignore both the meter and the tariff.
How to do it
- Identify the API product and model your application will actually call.
- Use the live API rate card's unit: input characters for text-to-speech, audio time or generations for other products.
- For current text-to-speech estimates, compare Flash/Turbo at $0.05 per 1,000 characters with Multilingual v2/v3 at $0.10 per 1,000 characters.
- Check your plan's included usage and model-specific concurrency limits separately.
- Monitor real usage and recalculate against the live rate card before changing plans or enabling paid usage.
Key facts
- API requests made with an API key deduct credits from the account.
- The public API rate card expresses usage in dollars and uses product-specific units.
- As of 2026-08-09, text-to-speech is listed at $0.05 per 1,000 characters for Flash/Turbo and $0.10 for Multilingual v2/v3.
- There is no universal API conversion from credits to finished-audio minutes.
- Plan level affects included usage and concurrency; model choice and input length affect text-to-speech cost.
Open the current official API pricing, free quota and rate-limit documentation before deploying.
▶ The 60-second explainer (script)
How does ElevenLabs API pricing work? Separate the account balance from the API rate card. Requests made with your API key deduct usage from your account, but the public API page prices each product and model in its own unit. As of August 9, 2026, text-to-speech lists Flash and Turbo at five cents per thousand input characters, while Multilingual v2 and v3 are ten cents per thousand. Transcription, music, dubbing and audio tools use audio hours, minutes or generations instead. Your plan also controls included usage and how many requests can run at once. Estimate from the exact model, product and live rate card; do not convert every thousand credits into one minute.
What authoritative sources say
People also ask
Is API pricing identical to website generation?
No. API calls deduct usage from your account, but the API rate card applies model- and product-specific units and discounts. Estimate from the API page, not the website rate.
How do I lower my API costs?
Choose the appropriate model, shorten unnecessary input, and size the plan for both included usage and peak concurrency. Recheck the live API rate card before budgeting.
What limits my API throughput?
Concurrency, how many requests run at once. Higher plans allow more simultaneous requests, and Enterprise offers elevated limits.
How much does one minute of API audio cost?
There is no fixed answer. Text-to-speech is metered by input characters and speaking duration varies by script, language and speed; other API products use different units.
The same question, asked other ways
- How does ElevenLabs usage based pricing work?