Sonic 3.6 is now Generally Available! Try it on the Playground or via the API now!
sonic-3.6) is our fastest, most natural text-to-speech model, with native support for 44 languages. It follows your transcript faithfully, voices confirmation codes and heteronyms correctly without preprocessing, and stays expressive enough to carry a real conversation.
What’s new in Sonic 3.6?
- Human-like naturalness — Sonic 3.6 is the most natural and human-sounding TTS model. It sounds more natural, more expressive, and better paced across every language. It varies its pacing and intonation to match the emotional context of the transcript, without SSML tags or explicit instructions: pause length adapts to the context of the sentence, and in-transcript disfluencies (e.g. “uhm”, “hmm”) produce a natural “thinking” pace.
- Native speaker quality — Sonic 3.6 speaks 44 languages and pronounces words and phrases the way a local would.
- Hinglish support — Sonic 3.6 has expanded support for Hindi transcripts written in Latin script, as well as improved pronunciation of Indian names and places. See our dedicated guide for more details.
- Two new languages — Odia (
or) and Urdu (ur), with instant voice cloning support. See language support.
PVCs created on older models will work on Sonic 3.6.
How can I use it?
Continuous updates (Recommended)
Using the model IDsonic-3.6 in your API calls will automatically keep you up to date with the most recent stable snapshot of the model.
For testing upcoming changes to our APIs, use sonic-preview. It is a beta model, not intended for production usage, and can change without notice.
Dated snapshots
If ensuring consistent behavior is important for your use case, we recommend using a dated version in production. A dated snapshot never changes once released.Voice selection
Choosing voices that work best for your use case is key to getting the best performance out of Sonic 3.6. For voice agents, we’ve found stable, pleasant, yet also realistic voices work better for voice agents. Some voices you should try include:- Skylar -
en-USFemale (ID:db6b0ed5-d5d3-463d-ae85-518a07d3c2b4) - Daniel -
en-USMale (ID:47c38ca4-5f35-497b-b1a3-415245fb35e1) - Jacqueline -
en-USFemale (ID:9626c31c-bec5-4cca-baa8-f8ba9e84c8bc) - Gemma -
en-GBFemale (ID:62ae83ad-4f6a-430b-af41-a9bede9286ca) - Archie -
en-GBMale (ID:ef191366-f52f-447a-a398-ed8c0f2943a1)
Language support
Sonic 3.6 speaks 44 languages. Set thelanguage field on your TTS request to one of the codes below:
English (en) | French (fr) | German (de) | Spanish (es) |
Portuguese (pt) | Chinese (zh) | Japanese (ja) | Hindi (hi) |
Italian (it) | Korean (ko) | Dutch (nl) | Polish (pl) |
Russian (ru) | Swedish (sv) | Turkish (tr) | Tagalog (tl) |
Bulgarian (bg) | Romanian (ro) | Arabic (ar) | Czech (cs) |
Greek (el) | Finnish (fi) | Croatian (hr) | Malay (ms) |
Slovak (sk) | Danish (da) | Tamil (ta) | Ukrainian (uk) |
Hungarian (hu) | Norwegian (no) | Vietnamese (vi) | Bengali (bn) |
Thai (th) | Hebrew (he) | Georgian (ka) | Indonesian (id) |
Telugu (te) | Gujarati (gu) | Kannada (kn) | Malayalam (ml) |
Marathi (mr) | Punjabi (pa) | Odia (or) | Urdu (ur) |
Odia and Urdu are new in Sonic 3.6. Curated voices for both are in the Voice Library, or clone a voice in either language.
Sonic 3.6 now supports locale codes using the
locale field in the API, so pass en-GB if you want 05/04/2026 to be read as “the fifth of April, twenty twenty-six.”. For more information, see Advanced Capabilities.Switching from Sonic 3.5?
Note that Sonic 3.6 is fully backwards compatible with Sonic 3.5.Older Models
For information onsonic-3, sonic-2, sonic-turbo, and sonic, see our page on Older TTS Models.
Where to go next
Try it out online
No sign-up or code required
Start building
Guides and best practices