The voice AI firm has introduced a new generative audio tool, AssemblyAI-TTS-1, intended for use in high-scale voice agents. The model is available for $15 per 1 million characters and supports more than 4,000 voices with capabilities for instant voice cloning and a time to first audio of about 90 milliseconds.
The new system recorded a generation speed of 162 characters per second, which is faster than counterparts such as ElevenLabs’ Eleven v4 and Google’s Gemini 3.8 Flash TTS. In evaluations by Artificial Analysis, the model ranked No. 12 in the Provider Voice Arena and No. 10 in the Controlled Voice Arena, while scoring 81.7% on the Pronunciation Robustness benchmark to place 16th overall.
