Cartesia releases Sonic-3.6 with improved naturalness across 44 languages
iamaliveix · x · 2026-08-18
Cartesia has released Sonic-3.6, their latest TTS model, featuring a significant step change in naturalness across 44 languages, just three months after the Sonic-3.5 launch. The update includes fundamental model improvements based on developer feedback.
Sonic-3.6 currently ranks #1 on Artificial Analysis across both the provider and controlled voice streaming leaderboards, offering best-in-class voices and underlying model performance. The model is available in beta starting today.
Related event: Cartesia Releases Sonic-3.6, Topping Speech Leaderboards(4 posts)→
More from Multimodal
- RTX 3090 MiniMax H3 Test: Rendering 5-Second Mario Clip in 2 Hours — 4as · 2026-08-18
- WanGP adds MiniMax Music support: a 2-minute song in 5m20s, prompting is everything — cocktailpeanut · 2026-08-18
- MiniMax H3 Reference to Video Workflow Not Utilizing Full VRAM — Select_Question5 · 2026-08-18
- ComfyUI-MiniMax-H3-LongMedia: open-source node pack for long-form H3 video on consumer GPUs — Independent-Ear-3035 · 2026-08-18
- Workflow Refinement: Making a 6-Minute Star Trek Video with MiniMax H3 — Glad-Hat-5094 · 2026-08-18
- Pinokio adds 1-click launcher for Wan2GP: local AI video on 6GB+ VRAM — cocktailpeanut · 2026-08-18