FULL STORY

ElevenLabs Launches Eleven v4, Tops TTS Leaderboards

ElevenLabs released Eleven v4 and v4 Turbo on Sept 28, topping TTS leaderboards. Early developer tests praised the model's unprecedented voice realism.

2026-09-28 ~ 2026-09-29 · 2 episodes · 19 posts

Episode 1 · ElevenLabs Launches Eleven v4 and v4 Turbo, Topping TTS Arena and Pronunciation Benchmarks (2026-09-28, 17 posts)

ElevenLabs has released its next-generation text-to-speech models, Eleven v4 and the low-latency v4 Turbo, which the company describes as its fastest and most emotionally expressive release yet. Benchmarks from Artificial Analysis show Eleven v4 topping the Provider Voice TTS Arena with an Elo of 1,319 across 1,674 appearances, ahead of competitors like Cartesia; it also ranks first on a pronunciation robustness benchmark, with its Controlled Voice ranking also mentioned. Third-party user mati shared the evaluation on X, calling Eleven v4 "clearly number one" in the comparison.

Confirmed

  • ElevenLabs officially launched two models: Eleven v4, focused on emotional expression, and v4 Turbo, built for low latency (@NevFlynn)
  • First place in the Provider Voice Arena: Elo 1,319 over 1,674 appearances, leading Cartesia (@ArtificialAnlys)
  • Topped the pronunciation robustness benchmark, with language support expanded to 90+ (@ArtificialAnlys)
  • Priced at roughly 5x its Google competitor (@ArtificialAnlys)

Why it matters

  • Eleven v4's dual first-place finishes in the arena and pronunciation benchmarks cement its lead in the TTS space
  • But at around 5x the price of competitors, the performance-versus-cost trade-off becomes a key consideration for users choosing a provider

Episode 2 · Early Testers Praise ElevenLabs v4 for Unprecedented Voice Realism (2026-09-29, 2 posts)

Early hands-on testers, including chrisfirst and D3VAUX, praise ElevenLabs v4 for its dramatically improved and unprecedentedly realistic generated speech.