Which model will be the last one standing?
Tune in Thursday as we drop ten LLMs into Agent Battle Royale and stream the whole thing, with AI casters calling the action in ElevenLabs voices.
Hosted by @jjacky from @OpenRouter and @tadaspetra.
https://x.com/i/broadcasts/1OxwbnrErygJB
Introducing the Realtime Skill
Teach your coding assistant to build an agent using Eleven Realtime or move a browser app off OpenAI Realtime on the same SDK.
Install with
> npx skills add elevenlabs/skills
Translate your voice into 90+ languages, with automatic language detection using one of the 17,500+ library voices or your own Instant or Professional Voice Clone.
Use the eagerness control to tune each conversation from the lowest latency to higher-quality translation for
Introducing Eleven Realtime Translation, a live voice translation model.
Each person speaks their own language and hears the other in theirs, in a natural voice. Now in alpha for enterprises via ElevenAPI.
Based on results from Artificial Analysis, Eleven Realtime v1 Max scores 98.8% on Big Bench Audio and starts speaking in 0.57s.
No model on the leaderboard is both faster and more accurate.
If you are already on OpenAI's Realtime API, you just need to make two lines of change.
Point the WebSocket transport at our endpoint and set the model to eleven_realtime_v1_mini or eleven_realtime_v1_max.
Eleven Realtime is live for developers.
It's a new class of Speech to Speech models complete with speech recognition, voice activity detection, turn-taking, LLM, and text to speech, available via the Realtime API.
v1 Mini is built for natural conversations at low cost, while v1
