Cartesia's Sonic-3.6 Goes GA, Tops TTS Leaderboards

On August 28, Cartesia released its real-time TTS model Sonic-3.6 for general availability, immediately topping multiple speech synthesis leaderboards: first on Voice Arena's US English board with 1085 Elo (tied with Google DeepMind's Gemini 3.1 Flash TTS), and first on both Artificial Analysis' controlled and provider TTS boards. The company reported positive feedback since preview.

Confirmed

• Sonic-3.6 went GA on 08-28, ranking first on Voice Arena US English (1085 Elo, tied with Gemini 3.1 Flash TTS)
• First place on both Artificial Analysis TTS controlled and provider leaderboards
• The Cartesia team says it decided in January to stop tuning the existing paradigm and fully rebuild the architecture, surpassing its predecessor Sonic-3.5 within two months

Independent Testing

Creator @socialwithaayan compared Cartesia and ElevenLabs in a real scenario, having both announce the same flight notification containing flight number, PNR, time and gate details (e.g., EK 418 / W8R4U). Cartesia showed a clear edge in clarity and naturalness on such detail-heavy content.

Why it matters

The TTS field has long been dominated by players like ElevenLabs. Cartesia reached the top through an architectural rebuild rather than fine-tuning, and its strong performance on digit- and code-heavy announcements offers direct reference value for customer-service and travel-broadcast applications.

2026-08-28 ~ 2026-08-28 · 5 related posts

• Cartesia's Sonic-3.6 tops TTS leaderboard, beats ElevenLabs in real-world test — socialwithaayan · 2026-08-28
• Test: #1 TTS Model Cartesia vs. ElevenLabs on Flight Announcements — socialwithaayan · 2026-08-28
• Cartesia Sonic-3.6 tops Voice Arena US English TTS leaderboard — hey_abusiddik · 2026-08-28
• [source] Cartesia Releases Sonic-3.6 TTS, Toping AI Audio Benchmarks After Architecture Rebuild — HazyResearch · 2026-08-28

1 near-duplicate retellings: socialwithaayan