Cartesia's Sonic-3.6 tops TTS leaderboard, beats ElevenLabs in real-world test
socialwithaayan · x · 2026-08-28
Cartesia released Sonic-3.6, retaking #1 on the Artificial Analysis TTS leaderboard (both controlled and provider voice boards). The company says it bet on rebuilding from the architecture up rather than incrementally tuning, overtaking its own Sonic-3.5 just two months later: listeners preferred Sonic-3.6 in up to 93% of blind head-to-head tests across fifteen locales, and its lead over #3 is wider than the gap from #3 to #12.
@socialwithaayan ran a real-world comparison: the same flight announcement containing a flight number, PNR, time and gate code (EK 418 / W8R4U / 10:20 PM / Gate C14) tested on both Cartesia and ElevenLabs — the difference was clearly audible, and alphanumerics remain a classic TTS stress test.
Related event: Cartesia Sonic-3.6 Tops TTS Leaderboard, Beating ElevenLabs in Tests(2 posts)→
More from Models
- DeepSWE Eval: GLM-5.3 Flash and Fable 5 Performance — zainhas · 2026-08-28
- User Returns to ChatGPT to Escape Claude's Weird Vocabulary — evielync · 2026-08-28
- Introducing PhoneLLM: Open Source Voice Model with 1/3 Latency and 1/18 Cost of GPT-5.6 Terra — charles_irl · 2026-08-28
- Omni 1.1 Flash Released on LMSYS Chatbot Arena — PleasantSir9581 · 2026-08-28
- Quality Difference: Minimax bf16 vs Pruned/int8 — Dapper_Astronaut_603 · 2026-08-28
- DeepSeek V4 Pro scores 59.7 on Tuesday Work Index, up 10.6 from preview — echen · 2026-08-28