Google launches Gemini 3 Deep Think and expands AI Mode Personal Intelligence globally

Gemini 3 Deep Think is Google's flagship reasoning model, positioned for the hardest math, science, logic and multi-step problems and gated to AI Ultra subscribers. It extends Google's reasoning-model line in direct competition with OpenAI's o-series successors and Anthropic's Fable 5 on reasoning-heavy benchmarks.

The broader-reach play is Personal Intelligence in AI Mode, now expanded to roughly 200 countries and 98 languages with no subscription required — a deliberate scale move to put Gemini in front of as many users as possible and defend Search against AI-native rivals. Google also rolled out Gemini 3.5 Flash Live Translate, a real-time speech-to-speech translation model supporting many languages, which Demis Hassabis and Sundar Pichai both amplified.

Google AI Studio lead Logan Kilpatrick struck a triumphant note ("bullish on Gemini"), and the week's news reinforced Gemini's expanding surface area — including powering Apple's new Siri AI and landing in NotebookLM.

The strategic thread is ubiquity plus tiering: free global reach for AI Mode to build habit and data, premium Deep Think reasoning for paying Ultra subscribers. Combined with reported AI subscription price cuts (a competitive blow to OpenAI), Google is leaning on distribution and pricing as much as raw capability. Readers should watch independent reasoning benchmarks for Deep Think and whether the global AI Mode expansion meaningfully shifts Search usage.