Alphabet releases Gemini 3.8 Live voice AI models

Google has launched Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. These two voice models were built specifically for responsive AI agents that handle multi-step tasks using advanced reasoning. The system can execute tools and API calls during dialogue, process visual data in near real-time, and switch across 97 languages. Google is working with Salesforce, Genspark, and Lumeris to integrate the models into enterprise workflows.

Gemini 3.8 Live is now available through Google AI Studio and the Gemini API. In testing by Artificial Analysis, Gemini 3.8 Live Extended Thinking recorded an 82.6 rating on the Speech-to-Speech Quality Index. The company is directing the technology toward customer service, productivity tools, and business automation as adoption scales. The Gemini app currently counts 950 million monthly active users, while its models process 22 billion API tokens per minute.

The release follows strong second-quarter financial results for Alphabet. Revenue climbed 24% year over year to $119.8 billion, exceeding analyst projections of roughly $117 billion. Google Services generated $94.5 billion, up 15%, while Google Cloud surged 82% to $24.8 billion. Operating income for the cloud unit expanded to $8.8 billion from $2.8 billion a year earlier. Diluted earnings per share reached $9.11, up from $2.31, boosted by a $99 billion net gain on equity securities that contributed $6.26 per share.

To support AI infrastructure demand, management raised its 2026 capital expenditures outlook to between $195 billion and $205 billion, up from a previous band of $180 billion to $190 billion. Management noted that using third-party capacity could pressure short-term margins. Meanwhile, Alphabet shares were trading near $356.76, giving the firm a market valuation around $4.25 trillion. The stock has gained about 40% over the trailing 12 months, compared to a 16% rise for the S&P 500.