Block's Dessa Voice Model: Trained on Speech Alone, Runs on a Laptop
On October 23, 2026, Jack Dorsey announced that a small research lab inside his fintech company Block had
voice AI
GPT-Live-1 in the API: OpenAI's Full-Duplex Voice Model, Explained for Creators (September 2026)
OpenAI has opened GPT-Live-1, its full-duplex voice model, to all API developers at $0.05/min. Deep dive: the architecture shift, official benchmarks, real pricing math, and five workflows that just got cheaper.
Gemini 3.8 Live and Extended Thinking: Google’s Real-Time Voice AI Explained
Google’s new voice models think while they talk: where to try them, what the independent benchmarks say, and three experiments creators can run today.
Gemini 3.5 Live Translate: Fluid Real-Time Voice Translation by Google (2026)
Google launched Live Translate in Gemini 3.5 for instant, natural voice translation. The feature breaks language barriers in real-time and supports multiple languages including Arabic.
xAI 21 New Flagship Grok Voices 2026: Multilingual Voice Models for Developers
xAI 21 New Flagship Grok Voices 2026: Multilingual Voice Models for Developers
Mistral Robostral Navigate 2026: Single RGB Camera Robot Navigation Model Explained
Mistral Robostral Navigate 2026: Single RGB Camera Robot Navigation Model Explained
Gemma 4 Real-Time Voice AI by Hugging Face and Cerebras
The AI landscape has just taken a massive leap forward in conversational technology. Hugging Face, in partnership with Cerebras, has unveiled an open-source, modular architecture that enables Google D