Pass IndexThe State of AISign in

Nova Sonic

A foundation model for conversational speech understanding and generation in five languages that accepts speech as input and provides speech with text transcriptions as output, with bidirectional streaming for real-time multi-turn conversations.

audio → text + audio · made by Nova

$3.4$13.6per Mtok in / out
AWS Bedrock

Sold by 1 way

SellerLaneRate
AWS Bedrock 0 ways$3.4per Mtok in$13.6per Mtok out$3.4per Mtok in, audio$13.6per Mtok out, audio
AWS Bedrock cloudnot on the seller's list todaystandard$3.4per Mtok in$13.6per Mtok out$3.4per Mtok in, audio$13.6per Mtok out, audioaws.amazon.com · read 2026-08-24

About

Nova Sonic — a text and audio model from Nova.

It takes audio and returns text and audio. The catalogue files it under transcribe and speak.

Every current figure

Maker
Nova
Register
model
Takes
audio
Returns
text + audio
Licence
not read

Known as 2 names

Amazon Nova Sonicamazon.nova-sonic-v1:0