Google's Gemini 3.8 Live tops voice leaderboard at $1.38/hour

Google Deepmind's release of Gemini 3.8 Live and 3.8 Live Extended Thinking adds two more options to the growing field of speech-to-speech models for developers. The Extended Thinking variant's first-place ranking on the Artificial Analysis Speech-to-Speech Leaderboard, with 82.6 percent, places it ahead of OpenAI's latest GPT-Live-1 models on that measure, while the standard Gemini 3.8 Live is positioned for voice agents that can make API calls in the background, process visual input, and keep talking at the same time, with support for over 97 languages.

The pricing gap is the clearest differentiator. Google's rates of $0.005 per minute for audio input and $0.018 for output translate to about $1.38 per hour of voice conversation, compared with at least $3.00 per hour for OpenAI's GPT-Live-1 at $0.05 per minute. At the same time, the article notes that OpenAI's model should still deliver more natural conversations thanks to full duplex, which lets it listen and speak at the same time, and that judging from the demos it sounds better.

For developers choosing between the two, the trade-off appears to hinge on whether leaderboard performance and lower cost outweigh the more natural conversational feel attributed to OpenAI's model. The article frames this as Google once again optimizing for price over quality, though the actual weight of that trade-off will depend on how each model performs in real deployments.