ElevenLabs has unveiled Flash, its latest text-to-speech model designed for ultra-fast performance, capable of converting text to speech in just 75 milliseconds.
This positions Flash among the fastest AI voice models available today, making it ideal for real-time applications like conversational AI.
Although Flash prioritizes speed, it comes with a trade-off: its voices are less expressive compared to the slower Turbo models.
However, ElevenLabs believes that for most users, especially in real-time scenarios, the difference will be negligible. Blind tests have shown that Flash outperforms other ultra-low-latency models.
Flash is available in two versions: v2, which supports only English, and v2.5, which supports 32 languages.
Both versions are accessible via ElevenLabs’ Conversational AI platform or through API integration, with the same pricing model—one credit for every two characters processed.
