Bland AI Launches Speech v3, Topping Audio Realism Blind Benchmark
EXM7777 · x · 2026-08-05
Bland AI has introduced Bland Speech v3, a new text-to-speech model described by the company as the first "Human Speech Engine."
Trained on over 100 million real human conversations, the model ranks first on the third-party Audio Realism Bench with an Elo score of 1237. It outperforms major competitors including Grok TTS, GPT Realtime 2, Gemini 2.5 Pro TTS, and ElevenLabs. The company also showcased a use case where they used just 5 seconds of old audio footage to help a 49-year-old father who lost his voice to a stroke speak again.
More from Multimodal
- MiniMax H3 Local Test: Generates 480P Video on an RTX 4050 Laptop — PixWizardry · 2026-08-05
- Exploring Workflows for Consistent AI Character Identity and Body Swap — GooDroop · 2026-08-05
- Developer Creates 3D Coastal Lighthouse Scene Using GPT-5.6 and Three.js — techartist_ · 2026-08-05
- Krea Introduces FLUX 3: First Model with Action-Prediction for Real-World Interaction — angrypenguinPNG · 2026-08-05
- AI-Generated Short: Smoking at the Back Door of a Mini Market (Ep. 5) — Stunning_Act_1530 · 2026-08-05
- AI-Generated Short: Magic, Cloak & Dagger (Ep. 4) — Juponce · 2026-08-05