Google launches Gemini 3.8 Live voice models at $1.38/hour, undercutting OpenAI's GPT-Live-1

The Decoder · rss · 2026-09-16

Google DeepMind has released two new audio models for developers: Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. Both top the Artificial Analysis speech-to-speech leaderboard.

At $1.38 per hour of voice conversation, Google significantly undercuts OpenAI's GPT-Live-1. The catch: GPT-Live-1's full duplex architecture should still sound more natural in conversation.

Original post →

More from Models

Models channel →