Google launches Gemini 3.8 Live to take on OpenAI’s GPT-Live-1 at a fraction of the cost
What happened
Google Deepmind launched two new audio AI models for developers, Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. These models take the top spots on the Artificial Analysis speech-to-speech leaderboard, beating other voice AI systems in performance. The key selling point is their price: Google offers these models at $1.38 per hour of voice conversation, significantly cheaper than OpenAI’s GPT-Live-1, which remains pricier despite its more natural-sounding full duplex audio capabilities.
Why it matters
Google’s move breaks the current cost barrier for live voice AI applications. The $1.38 hourly rate places Gemini Live models in reach for developers and businesses that want to build interactive voice experiences without incurring heavy expenses. This pricing pressure could force OpenAI and others to rethink their voice service economics, especially as conversational AI proliferates into more practical real-time use cases like customer support, virtual assistants, and voice interfaces.
Lower cost access brings voice AI closer to mass adoption outside elite tech labs, enabling startups and small businesses to experiment with live speech-to-speech AI. Google also signals a focus on balancing top-tier performance with practical operational costs, which is crucial as voice AI expands into high-volume, real-world deployment.
What to watch next
The biggest factor to observe will be user experience differences between Google’s models and OpenAI’s GPT-Live-1. Google’s models do not operate in full duplex, meaning they may introduce slight delays or less fluid conversation flow compared to OpenAI’s approach, which supports simultaneous listening and speaking.
Developer adoption and ecosystem integration matter too. How Google rolls out API access, documentation, and usage flexibility will influence whether Gemini Live gains broad traction versus GPT-Live-1 or other speech AI competitors. Watch for adoption in call centers, educational tools, and real-time translation where cost and performance tradeoffs become concrete business decisions.
Finally, Google’s pricing invites competitive responses. Expect adjustments from Microsoft, OpenAI, and other cloud providers aiming to protect or wrest market share by improving price-performance ratios in live voice AI.
AI Quick Briefs Editorial Desk