Loading video...
Video Failed to Load
Meet Flash. Our newest model that generates speech in 75ms + application & network latency. You’ve never experienced human-like TTS this fast.
173,598 views • 1 year ago •via X (Twitter)
11 Comments

Flash is our recommended model for low-latency, conversational voice agents. You can use Flash today in our Conversational AI platform Or build directly via the API using model id “eleven_flash_v2” and “eleven_flash_v2_5”:

Flash v2 is English only and Flash v2.5 supports 32 languages They both cost 1 credit for every 2 characters

It has a slightly lower quality and emotional depth that the Turbo models but significantly lower latency. And the Flash model quality is still higher than competitor models. Check out our guide on models to find the best for your use case:

Hear from @maxilevi__, one of the developers who lead the engineering work behind this update:

Introducing Vehrbal, the AI that converts audio into SOAP notes! Say goodbye to wasted time and hello to effortless note-taking. Experience the power of fast, simple, and efficient with Vehrbal today.

From 250ms to 75ms 🏎️🏎️ Great work @maxilevi__ and team!

incredible product but the price just doesn't make sense for anything at scale. openai ttl can be run at about 1/10th of the price

Your models are great! I just need better pricing. I am building a voice intensive app. Can we come into an agreement? I am a paid customer.

holy shit

This combined with conversational agents is a magical experience @maxilevi__ @Marko_Jozef

man, if you are a kid today. this is the golden era for you.
