Google Gemini 3.8 Live undercuts OpenAI on voice AI pricing
Google Deepmind's release of Gemini 3.8 Live and 3.8 Live Extended Thinking marks a significant pricing shift in the real-time voice AI market. At $1.38 per hour, these models undercut OpenAI's GPT-Live-1 substantially while achieving top leaderboard rankings on speech-to-speech benchmarks. The trade-off centers on full-duplex capability, where OpenAI's offering maintains a naturalness edge. This move signals intensifying competition in conversational AI infrastructure, forcing developers to weigh cost efficiency against interaction quality. For teams building voice-first applications, the pricing delta reshapes deployment economics.
Modelwire context
Analyst takeThe full-duplex gap is the number that actually matters here, and it's conspicuously unquantified. Artificial Analysis ranks Gemini 3.8 Live at the top on speech-to-speech benchmarks, but 'naturalness' in interruption handling and turn-taking is the friction point enterprise voice deployments hit first, and no hard metric for that gap has been published.
The pricing move lands the same day TechCrunch reported that U.S. data centers could consume more natural gas than Germany and Japan combined by 2035. That story framed energy efficiency as an existential business constraint, not a sustainability checkbox. Google's $1.38 per hour price point only holds as a competitive weapon if the underlying inference cost structure is sustainable at scale, and that is a harder question once energy pricing volatility enters the model. The two stories together suggest that the race to undercut on voice AI pricing may be running directly into a cost floor that neither Google nor OpenAI fully controls.
Watch whether OpenAI responds with a GPT-Live-1 price cut within 60 days. If they hold price and instead publish latency or naturalness benchmarks, that signals they're betting the full-duplex quality gap is wide enough to defend margin rather than compete on cost.
Coverage we drew on
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsGoogle Deepmind · Gemini 3.8 Live · Gemini 3.8 Live Extended Thinking · OpenAI · GPT-Live-1 · Artificial Analysis
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. The Decoder originally reported this story as “Google launches Gemini 3.8 Live to take on OpenAI's GPT-Live-1 at a fraction of the cost”. The full content lives on the-decoder.com. If you’re a publisher and want a different summarization policy for your work, see our takedown page.