Best AI model to generate voice from text
Narration, voiceover, dubbing and reading text aloud. There is no open voice quality leaderboard in the sources checked, so the list is ordered by price, not quality, with an open weights alternative.
Suggested models
Note: this list is ordered by price, not quality. There is no open quality leaderboard in the sources checked when you want to generate voice from text, so it takes one main model per company, from the lowest to the highest price per thousand characters on each company's official page.
1 Azure Neural TTS Microsoft
Why it fits: Microsoft's main model to generate voice from text. According to the official page: Languages: over 140 languages and variants, with over 400 voices · Voice cloning: yes, with limited access · Latency: not stated (official page)
Ordered by price, not by quality.
- Price: US$ 15 per 1M characters (that's US$ 0.015 per 1,000 characters) (source: Microsoft, official page; standard neural voices; official Azure price list)?what it is
- Estimated cost (see the How we calculate block): about US$ 0.15 (estimate)
- Weights: Undisclosed?
- Alternative: open weights: Qwen3-TTS · cheaper: Google Cloud TTS WaveNet, US$ 0.004 per 1,000 characters
2 Grok Text to Speech xAI
Why it fits: xAI's main model to generate voice from text. According to the official page: Languages: 20, including Brazilian Portuguese · Voice cloning: yes, from a short clip · Latency: streaming with latency tuning (official page)
Ordered by price, not by quality.
- Price: US$ 15 per 1M characters (that's US$ 0.015 per 1,000 characters) (source: xAI, official page)?
- Estimated cost (see the How we calculate block): about US$ 0.15 (estimate)
- Weights: Undisclosed?
- Alternative: open weights: Qwen3-TTS · cheaper: Google Cloud TTS WaveNet, US$ 0.004 per 1,000 characters
3 Google Cloud TTS Chirp 3: HD Google
Why it fits: Google's main model to generate voice from text. According to the official page: Languages: over 50 variants, including Brazilian Portuguese · Voice cloning: yes, with the Instant Custom Voice feature · Latency: low, with text streaming (official page)
Ordered by price, not by quality.
- Price: US$ 30 per 1M characters (that's US$ 0.03 per 1,000 characters) (source: Google Cloud, official page)?
- Estimated cost (see the How we calculate block): about US$ 0.30 (estimate)
- Weights: Undisclosed?
- Alternative: open weights: Qwen3-TTS · cheaper: Google Cloud TTS WaveNet, US$ 0.004 per 1,000 characters
4 OpenAI TTS HD (tts-1-hd) OpenAI
Why it fits: OpenAI's main model to generate voice from text. According to the official page: Languages: the same as Whisper, including Portuguese; voices tuned for English · Voice cloning: not stated · Latency: tts-1 is faster; HD favors quality · Also: OpenAI has said this model will be shut down (official page)
Ordered by price, not by quality.
- Price: US$ 30 per 1M characters (that's US$ 0.03 per 1,000 characters) (source: OpenAI, official page)?
- Estimated cost (see the How we calculate block): about US$ 0.30 (estimate)
- Weights: Undisclosed?
- Alternative: open weights: Qwen3-TTS · cheaper: Google Cloud TTS WaveNet, US$ 0.004 per 1,000 characters
5 ElevenLabs v4 ElevenLabs
Why it fits: ElevenLabs's main model to generate voice from text. According to the official page: Languages: over 90 · Voice cloning: yes, high fidelity · Latency: not stated · Also: up to 10,000 characters per request (official page)
Ordered by price, not by quality.
- Price: US$ 0.08 per 1,000 characters (source: ElevenLabs, official page)?
- Estimated cost (see the How we calculate block): about US$ 0.80 (estimate)
- Weights: Undisclosed?
- Alternative: open weights: Qwen3-TTS · cheaper: Google Cloud TTS WaveNet, US$ 0.004 per 1,000 characters
How we calculate
Change the numbers and the cost of every model on the page is recalculated right away. Without JavaScript, the default values below apply.
Quantity = characters of text to narrate ÷ 1,000, in thousand characters. Cost = quantity × price per 1,000 characters.
With these numbers: 10 thousand characters (estimate).
Cost of each model = quantity × the price per thousand characters published by the company. When the company uses another unit, the conversion is shown next to the price. It is a reference calculation and does not change the order of the list.
Other paid models we checked
Other versions and companies, with the price from each company's official page. They are left out of the list above when another model from the same company is already among the first five, when they do not meet this page's criterion, or when the company does not publish a price per thousand characters.
| Model | Price and source | Estimated cost |
|---|---|---|
| Google Cloud TTS WaveNet Google | US$ 4 per 1M characters (that's US$ 0.004 per 1,000 characters) (source: Google Cloud, official page) | about US$ 0.04 (estimate) |
| OpenAI TTS (tts-1) OpenAI | US$ 15 per 1M characters (that's US$ 0.015 per 1,000 characters) (source: OpenAI, official page) | about US$ 0.15 (estimate) |
| ElevenLabs Flash / Turbo ElevenLabs | US$ 0.04 per 1,000 characters (source: ElevenLabs, official page) | about US$ 0.40 (estimate) |
| MiniMax Speech 2.8 Turbo MiniMax | US$ 60 per 1M characters (that's US$ 0.06 per 1,000 characters) (source: MiniMax, official page) | about US$ 0.60 (estimate) |
| MiniMax Speech 2.8 HD MiniMax | US$ 100 per 1M characters (that's US$ 0.10 per 1,000 characters) (source: MiniMax, official page) | about US$ 1.00 (estimate) |
| gpt-4o-mini-tts OpenAI | price not published (source: OpenAI, official page; billed by tokens: US$ 0.60 per 1M input text and US$ 12 per 1M output audio; no per character price) | no estimate: the company does not publish a price per thousand characters |
FAQ
What is the best AI model to generate voice from text?
By this page's rule, the list starts with Azure Neural TTS, Grok Text to Speech, Google Cloud TTS Chirp 3: HD, OpenAI TTS HD (tts-1-hd), ElevenLabs v4. Note: this list is ordered by price, not quality. There is no open quality leaderboard in the sources checked when you want to generate voice from text, so it takes one main model per company, from the lowest to the highest price per thousand characters on each company's official page.
How much does it cost to generate voice from text with AI?
With the default values in the How we calculate block, the estimate ranges from US$ 0.15 to US$ 0.80 among the listed models with a published price. It is only an estimate and you can adjust the numbers on the page.
Is there an open weights option to generate voice from text?
Yes: Qwen3-TTS.
Does this page use AI to recommend?
No. The recommendation comes from fixed rules over public data from Epoch AI, OpenRouter, Hugging Face and each company's official pricing page, in the latest version we checked. The typed text is only compared with keywords.
Terms used on this page
Use cases with their own page
Official pricing pages checked
- Microsoft, latest version we checked.
- Google Cloud, latest version we checked.
- OpenAI, latest version we checked.
- ElevenLabs, latest version we checked.
- MiniMax, latest version we checked.
- xAI, latest version we checked.
Sources
- Hugging Face Hub, metadados públicos do Hub; licença de cada modelo no repositório, latest version we checked.
Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.