Text-to-speech price per million characters: 11 rates vs Sume
Eleven TTS rates converted to dollars per 1M characters, from $11 to $150. Sume TTS 1.0 is $47.50. Each is read from the vendor page, and caveats are listed.

Sume TTS 1.0 costs $47.50 per million characters, which puts it eighth of the eleven rates below: more expensive than Rime Mist v3 at $30 and Deepgram Aura-2 at $30, and slightly cheaper than Rime Coda at $50. All figures come from the vendor pages read on 2026-10-05 and are converted to dollars per million characters so they line up. Voice quality, latency and language coverage are not in the price.
The ranking
Per-1K rates are multiplied by 1,000 to get a per-million figure. Where two vendors list the same rate for the same model name, the table shows them on one row. Hume's per-character rates are listed per plan, so the rows show the lowest and highest.
| Rank | Model | Listed unit | Per 1M characters |
|---|---|---|---|
| 1 | Speechmatics TTS | $0.011/1K characters, first 1M free | $11.00 |
| 2 | Cloudflare Aura-1 / Deepgram Aura-1 | $0.015/1K | $15.00 |
| 3 | Microsoft MAI-Voice-2.1-Flash | $15/1M | $15.00 |
| 4 | Microsoft MAI-Voice-2.1 | $22/1M | $22.00 |
| 5 | Deepgram Aura-2 / Cloudflare Aura-2 | $0.030/1K | $30.00 |
| 6 | Rime Mist v3 | $0.03/1K | $30.00 |
| 7 | Deepgram Flux TTS | $0.045/1K | $45.00 |
| 8 | Sume TTS 1.0 | $0.0475/1K | $47.50 |
| 9 | Rime Coda | $0.05/1K | $50.00 |
| 10 | Hume Business plan rate | $0.05/1K | $50.00 |
| 11 | Hume Free/Starter/Creator rate | $0.15/1K | $150.00 |
Read these before quoting
Speechmatics lists its first 1M characters as free, so the first million costs nothing there. Microsoft's two MAI-Voice rates are from the launch announcement and are stated per million characters. Hume's rates depend on plan, and its cheapest per-character figures come from larger monthly plans, so a small user pays the top of the range.
Soniox's TTS is priced per hour of generated speech at $0.70, so it cannot be placed on a per-character list without assuming a speaking rate, and it is left out.
Sume's rate and what it means
Sume TTS 1.0 is the provider list of $0.038 per 1K characters times 1.25, which is $0.0475, billed per request up to 20,000 characters (a full request is $0.95). Spaces and punctuation count as characters. Check the current figure on the API pricing page.
- Against the cheapest rate in the list ($11): about 4.3 times.
- Against the $30 group: $17.50 more per million.
- Against Flux TTS ($45): $2.50 more per million.
Budgets at three volumes
Per-million prices are easy to misjudge, so here are four of the rates at three monthly volumes. Each cell is characters in millions times the per-million rate, with no free allowances, plan fees or taxes applied. Speechmatics is left out because its first million characters are free, which changes the shape of the curve.
| Characters per month | Rime Mist v3 | Deepgram Aura-2 | Sume TTS 1.0 | Rime Coda |
|---|---|---|---|---|
| 1M | $30.00 | $30.00 | $47.50 | $50.00 |
| 10M | $300.00 | $300.00 | $475.00 | $500.00 |
| 100M | $3,000.00 | $3,000.00 | $4,750.00 | $5,000.00 |
A practical way to use the list
Take your monthly character count, multiply by the rate in the last column and divide by one million. At 5 million characters a month, Rime Mist v3 is $150, Sume is $237.50 and Rime Coda is $250. The spread between the cheapest and the most expensive entry on the list is $139 per million, so it matters far more than the $2.50 difference between neighbours at the Sume end.
Decide on voice first, then use this list to see what the voice you chose costs on each service that offers it. Most voices are exclusive to one vendor, so the real choice is often between two or three comparable voices, and in that case the per-million gap is a tiebreaker, not the main reason. Budget with the rate you will actually pay, not the cheapest one on the page.
Sources
Rates are from Speechmatics, Deepgram, Cloudflare, Microsoft AI, Rime and Hume. Re-read each before you use a number in a budget.
Sources
Related posts
More in Comparisons
- 3-hour hearing transcript: MAI Streaming vs Sume STT speaker labels
A 180-minute hearing costs about $1.62 on MAI-Transcribe-2-Streaming and $1.81 on Sume STT with the split, but Sume's result has no speaker labels.
- Wan 3.0, H3 and H3 Max tie at $0.0625 a second at 480p on Sume
At 480p, wan-3.0, minimax-h3 and minimax-h3-max all bill $0.0625 a second on Sume, $0.63 for 10 seconds. Break the tie on length and tiers above.
- Does TikTok auto-label AI video? Only its own AI effects
TikTok labels content made with its official AI effects automatically. A Sume clip is made outside the app, so you disclose it yourself when it looks real.
- Transcribe 500 hours: Speechmatics tiers vs Sume STT cost
500 hours of audio: $60 on Speechmatics Melia 1, $115 Standard, $190 Enhanced, $300 on Sume STT 1.0. The math, the 10-minute job limit, free credit.
Written by Sume