Bland Speech v3 Launches: Beats ElevenLabs & OpenAI, Restores Stroke Patient's Voice

LinusEkenstam · x · 2026-08-05

Bland has launched Speech v3, introduced as the world's first Human Speech Engine. The model was trained on over 100 million real human conversations.

In Design Arena’s Audio Realism benchmark, Speech v3 ranks at the top, outperforming major models including ElevenLabs, Grok, Cartesia, and OpenAI. To demonstrate its capabilities, the team used just 5 seconds of old footage to help James, a 49-year-old father who recently suffered a stroke, successfully regain his voice.

Related event: Bland AI Launches Speech v3 Voice Engine(3 posts)→

Original post →

More from Multimodal

Multimodal channel →