MAI-Voice-2.1 Hindi: six voices listed, and a Hindi request on Sume
Microsoft's MAI-Voice-2.1 page lists six hi-IN voices with different style sets. Compare that with a Hindi request on Sume TTS and what the language field does.

Microsoft's Learn page for MAI-Voice-2.1 and MAI-Voice-2.1-Flash lists six Hindi (hi-IN) voices: Arjun, Dhruv, Grant, Harper, Kavya and Priya (read 2026-10-05). They are not interchangeable. Arjun, Dhruv, Kavya and Priya list long emotion style sets (joyful, sad, surprised and more), and Dhruv, Kavya and Priya add shouting, softvoice and whispering. Grant lists only audiobook and neutral. Harper lists agent, customer_call_center, educational, narrator and neutral.
Pick by what the script does
- Announcements and explainers: a voice with narrator or educational.
- Support lines and IVR scripts: customer_call_center or agent.
- Story or ad reads: one of the voices with joyful, excited or whispering.
The same job on Sume
Sume TTS takes the language per request. For Hindi you send language: "hi" with a voice that speaks Hindi. Sume's voice library tags every voice with one of 16 languages, and hi is one of them, alongside en, ko, ja, zh, es, fr, de, pt, it, nl, pl, ru, sv, tr and tl. If a voice's language and the request disagree, Sume answers 409 tts_voice_language_mismatch before any job or charge, unless you send confirm_language_mismatch: true to override it.
Price for a Hindi script
| Service | Price | 500-character Hindi line |
|---|---|---|
| MAI-Voice-2.1 | $22 per 1M characters | $0.011 |
| MAI-Voice-2.1-Flash | $15 per 1M characters | $0.0075 |
| Sume TTS 1.0 | $0.0475 per 1,000 characters | $0.02375, billed to 3 cents |
Test before you commit
Write one 500-character Hindi paragraph with a price, a date and an English brand name. Render it with two voices on each service, play the files to a native speaker, and keep notes on numbers and the brand name, which is where mixed-script text usually slips. On Sume the renders cost a few cents. Repeat the test when a vendor ships a new version, since voice behavior can shift between releases.
What the table leaves out
Sume's cent rounding applies per job, so many short lines cost more than one long job. MAI-Voice-2.1 is a public preview without an SLA on Microsoft's page, and its cloning is gated. Sume's price includes its own margin over the provider list. The MAI prices come from Microsoft's model page, which lists $22 and $15 per 1M characters; the Learn page itself only points to Azure's pricing page, so confirm before you budget.
Sources
Related posts
More in Models
- Mandarin Chinese text to speech API: MAI-Voice-2.1 zh-CN vs Sume zh
MAI-Voice-2.1 lists Chinese (Simplified) as zh-CN; Sume tags zh and bills by character. Audition Mandarin ad copy for 1 cent and know what to check.
- Which Foundry models add C2PA and a watermark: GPT Image, FLUX.2, MAI
Microsoft's Foundry page lists the image, audio and voice models that get C2PA credentials and watermarks. What that does and does not tell you about Sume.
- MiniMax H3 API: the cheapest Sume job is 5 s at 480p, $0.3125
On Sume the minimax-h3 model takes 5 to 15 seconds at native 480p or 768p, with 2K and 4K upscales billed on request. Price table by length inside.
- MiniMax H3 Max: five free fal runs a day, then the Sume cost per clip
fal's page offers five free H3 Max generations a day of up to 15 seconds. Use them to test a prompt, then price the paid clip on Sume at 480p, 768p or 1080p.
Written by Sume