Text to speech, voice cloning and music tools price by characters, credits, minutes or songs, so their prices rarely line up. This page lists what each tool charges, straight from its own pricing page, and works out a common unit where the vendor publishes enough to do the math: dollars per 1 million characters for voice, dollars per track or per download for music. Official vendors come first. Platforms that resell other companies' models are listed separately below.
This page has no affiliate links. If that changes, each affiliate link will be labeled where it appears and named on the Disclosure page. A commission would never change which tools are listed, their order or the prices shown.
On the official pages checked for this site, text to speech APIs range from $4 per 1 million characters (Amazon Polly Standard and Google Cloud Standard voices) to $160 per 1 million characters (Google Cloud Studio voices). ElevenLabs' API lists its v4, v3 and v2 Multilingual voices at $0.08 per 1,000 characters, which is $80 per 1 million. For music, Google's Lyria 3.5 API is $0.08 per song and Stable Audio 3.0 is 8 credits ($0.08) per generation. Suno sells plans with credits and a monthly download cap, and does not state how many credits a song uses.
| Tool | What you buy | Per 1M characters | Source |
|---|---|---|---|
| Amazon Polly | Standard · Neural · Generative · Long-Form voices, API | $4.00 · $16.00 · $30 · $100.00 | aws.amazon.com/polly/pricing, checked November 9, 2026 |
| ElevenLabs | API: v4, v3, v2 Multilingual · Flash / Turbo, v4 Turbo, v3 Conversational. Worked out: $0.08 or $0.04 per 1K characters x 1,000 | $80 · $40 | elevenlabs.io/pricing/api, checked November 9, 2026 |
| Fish Audio | API, S2.1 Pro. Billed per 1M UTF-8 bytes, not characters (see note) | $15 per 1M UTF-8 bytes | fish.audio/developers, checked November 9, 2026 |
| Google Cloud Text-to-Speech | Standard · WaveNet · Neural2 · Chirp 3: HD · Studio voices | $4 · $4 · $16 · $30 · $160 | cloud.google.com/text-to-speech/pricing, checked November 9, 2026 |
| Inworld AI | TTS-2 · TTS-2 Flash, On-Demand (pay as you go) | $25 · $15 | inworld.ai/pricing, checked November 9, 2026 |
| MiniMax | API, speech-2.8-hd · speech-2.8-turbo | $100 · $60 | MiniMax API pricing, checked November 9, 2026 |
| OpenAI | API, tts-1 · tts-1-hd | $15.00 · $30.00 | OpenAI API pricing, checked November 9, 2026 |
Subscription apps sell credits or minutes, not characters. Worked out at full use of the plan allowance.
Spot a price that changed? Report the change. Prices are updated only when the new figure can be read on the vendor's own page.
| Tool | What you buy | Price | Source |
|---|---|---|---|
| AIVA | Pro plan, billed annually, 300 downloads a month. Worked out: €33 / 300 | €33 a month + VAT; about €0.11 per download | aiva.ai/pricing, checked November 9, 2026 |
| ElevenLabs | Music API | $0.15 per minute (a 3-minute track is $0.45, worked out) | elevenlabs.io/pricing/api, checked November 9, 2026 |
| Gemini API, Lyria 3.5 (full song) · Lyria 3 Clip Preview (30s) | $0.08 per song · $0.04 per song | Gemini API pricing, checked November 9, 2026 | |
| Stability AI | API, Stable Audio 3.0 (up to six minutes). Worked out: 8 credits x $0.01 | $0.08 per generation | Stability AI pricing, checked November 9, 2026 |
| Suno | Pro plan, $10 a month, 20 song downloads a month. Worked out: $10 / 20 | $0.50 per download at full use | suno.com/pricing, checked November 9, 2026 |
Spot a price that changed? Report the change. Prices are updated only when the new figure can be read on the vendor's own page.
Third-party platforms that sell access to other companies' voice and music models at their own prices. A reseller's price is not the model maker's price, and being listed here does not mean the model maker works with or approves of the reseller.
Spot a price that changed? Report the change. Prices are updated only when the new figure can be read on the vendor's own page.
Plain dated observations. A gap is never a judgment about a tool.
On the official pages checked here, Amazon Polly Standard and Google Cloud Standard voices are both $4 per 1 million characters. Each vendor sells several voice tiers at different prices, and voice quality is not compared on this site.
It depends on characters per minute, which vendors measure differently. Amazon's own example puts 1 million characters at about 23 hours of speech, so Polly Neural at $16 per 1 million works out to about $0.69 an hour. ElevenLabs' API page pairs $0.08 per 1,000 characters with about $0.08 a minute, about $4.80 an hour.
Not on its pricing page. It lists 2,500 credits a month on Pro and 10,000 on Premier.
Not always. fal lists ElevenLabs v2 at $0.1 per 1,000 characters; ElevenLabs' own API lists $0.08.