The global technology company Google has announced a major expansion of its artificial intelligence portfolio, headlined by the global release of Gemini 3.5 Flash. As the first release in the new Gemini 3.5 family, Flash is optimized for high-speed, long-horizon agentic workflows and coding, claiming a fourfold output speed advantage over comparable models. It is now live in the consumer Gemini app and Google Search’s AI Mode, and will soon power “Gemini Spark,” a 24/7 personal assistant debuting in Beta next week. A larger flagship model, Gemini 3.5 Pro, is currently in internal testing for release next month.
Alongside the foundational model updates, Google introduced Co-Scientist, a specialized multi-agent AI system published in Nature (a British weekly international scientific journal publishing peer-reviewed research across the natural sciences) built on the Gemini model designed to generate, debate, and refine scientific hypotheses. Integrated with data tools like AlphaFold, it is being tested by institutions like MIT and Stanford to accelerate research into diseases and cellular aging. Individual researchers can access this via an experimental “Hypothesis Generation” tool within Gemini for Science.
Concurrently, Google AI Studio (a browser-based prototyping environment designed to help developers and creators experiment with Google’s generative AI models) received significant upgrades, allowing developers to build native Android apps directly from natural language prompts using Kotlin code.
