ElevenLabs Eleven v4 & v4 Turbo: Emotive, Multilingual Text-to-Speech Models

Melvin Vivas · X video post · 2026-09-29 · 2:36 · 849 views · Open on X

Topics: Industry Trends & Job Market, LLM Fundamentals · Level: beginner

Summary

Melvin Vivas reshares ElevenLabs' launch of Eleven v4 and Eleven v4 Turbo, which ElevenLabs calls its fastest and most emotive voice models so far. ElevenLabs says the models are ranked #1 by Artificial Analysis. The demo reel is unedited audio generated from prompts. It shows emotional acting, multi-speaker banter, better Professional Voice Clones, and a low-latency voice-agent call that switches to Mandarin partway through. It's a product announcement, not a tutorial, but it's useful for keeping up with the state of voice AI.

Key points

• Two models launched: Eleven v4, the expressive speech model, and Eleven v4 Turbo, an ultra-low-latency version aimed at real-time and conversational use.
• ElevenLabs says the models are ranked #1 by Artificial Analysis, an independent benchmarking site. That's worth checking if you're comparing TTS providers.
• The demo says all audio came straight from the prompts shown, with no edits. It includes acting with emotion, hesitation, interruptions and multiple speakers.
• Professional Voice Clones are claimed to reach 'significantly better speaker similarity' with v4.
• Claimed features: the voice stays consistent across text of any length, it works natively in 100 languages, and Turbo has ultra-low latency.
• The voice-agent demo is a healthcare prior-authorization phone call. Partway through, the caller asks for medication information in Mandarin and the agent switches languages. This shows a realistic multilingual voice-agent use case. In the machine-generated transcript, the Mandarin section came out garbled.
• Takeaway for engineers: pick the expressive model for content and narration, and the Turbo or low-latency model for real-time voice agents where response time matters.

Resources mentioned

• ElevenLabs · tool · elevenlabs.io · free
  Voice AI platform for speech synthesis and speech-to-text, used here in a real-time voice sentiment analysis demo.
  Also in: A Cheaper Voice Option Than ElevenLabs, Credited to @thorwebdev (Melvin Vivas on X · notes), Demo: Real-Time Voice Sentiment Analysis with Jev and ElevenLabs (Melvin Vivas on X · notes), OpenAI's Agentic Software Factory: Build-Review-Deploy-Observe Loop (Melvin Vivas on X · notes)
• Eleven v4 · tool · elevenlabs.io · check price
  ElevenLabs' new expressive speech model, which the company calls its most emotive. Supports 100 languages and keeps the voice consistent over long text.
• Eleven v4 Turbo · tool · elevenlabs.io · check price
  Ultra-low-latency version of Eleven v4 for real-time voice agents and conversational apps.
• ElevenLabs Professional Voice Cloning · tool · elevenlabs.io · paid
  ElevenLabs feature for building high-fidelity clones of a specific voice.
• Artificial Analysis · website · artificialanalysis.ai · free
  Independent benchmarking site that compares AI models and providers, including text-to-speech, on quality, speed and price.
  Also in: Be Skeptical of Model Leaderboards: Muse Spark vs Astra (Melvin Vivas on X · notes), Pipette: Open-Source Benchmarking for On-Device Models (Melvin Vivas on X · notes), GLM 5.2 Hits 446 tok/s on Fireworks AI (Melvin Vivas on X · notes), GLM-5.2 on Fireworks: Top Open-Weights Model on GDPval-AA (Melvin Vivas on X · notes) and 1 more