Drama 3 tops pronunciation robustness at 94.6%, beating all proprietary TTS
ArtificialAnlys · x · 2026-10-29
Fish Audio Drama 3 scores 94.6% overall on the Artificial Analysis Pronunciation Robustness benchmark, ranking #1 of all models tested, open weights or proprietary. Human reviewers judge whether TTS models correctly pronounce challenging text across four categories. Results: Expanding Shorthand 96.6% (#1); Contextually Appropriate 96.5% (#2, behind Gemini 3.8 Flash TTS at 97.9%); Preserving Exact Sequences 87.1% (#3, behind HeyGen Voice Turbo at 91.8%); Standalone Terms 91.8% (#13, behind Drama 3 Preview at 95.9%).
Related event: Fish Audio open-sources Drama 3, tops open-weight TTS rankings(5 posts)→
More from Multimodal
- Native 3K MiniMax H3 video on a 4090: one ComfyUI node kills the VRAM crash — friedlc · 2026-10-30
- 'Draft-then-upscale' will likely become the default image workflow soon — socialwithaayan · 2026-10-30
- Inworld launches Reasoning for Realtime TTS-2, first context-steering TTS model — Scobleizer · 2026-10-30
- Runway's App Studio turns short films into full video games, no coding required — tlakomy · 2026-10-30
- Phosphene generates SOTA video clips fully locally on Apple Silicon Macs — cocktailpeanut · 2026-10-30
- ElevenLabs launches SFX v3: richer, more precise sound effects from a single prompt — umesh_ai · 2026-10-30