Google says Gemini in Google Sheets reached 70.48% on the full SpreadsheetBench benchmark, approaching human expert ability. The company attributes the result to product-specific tuning plus stronger verbalization and coding behavior inside Sheets.
Google DeepMind said Gemini 3.1 Flash-Lite is rolling out in preview through the Gemini API and Google AI Studio. The company positioned it as the most cost-efficient Gemini 3 model, with lower price, faster performance, and tunable thinking levels.
Google AI used X on March 6, 2026 to direct developers to Nano Banana 2, saying the model is available through the Gemini API in Google AI Studio and Vertex AI. Google’s linked post positions Nano Banana 2, or Gemini 3.1 Flash Image, as a high-quality and faster image model designed for real application workloads.
Google said on February 24, 2026 that it is rolling out a new agent step in Opal for all users. The feature lets Opal choose the right tools and models for a goal, adds Memory across sessions, and pushes the product from static workflow wiring toward more interactive agentic workflows.
Google used AI Impact Summit 2026 to bundle infrastructure, public-sector grants, and workforce programs into one AI adoption package. The announcements include $15 billion for Indian cloud and AI infrastructure, two new $30 million initiatives for government and science, and a training effort aimed at 720,000 workers.
Google's March Pixel Drop begins rolling out over the next several weeks with Gemini actions across apps, multi-object Circle to Search, Magic Cue restaurant suggestions and new Pixel Watch safety features. The update also expands scam detection, call notes, Find Hub and Satellite SOS availability.
Google opened the AI Center Berlin on March 5 and said the site will connect teams from Google DeepMind, Google Research and Google Cloud with researchers, businesses and policymakers. At the launch, Google also announced long-term research partnerships with TUM and Helmholtz Munich.
Google and Taiwan's NHIA said they are combining 20+ years of health data, an AI-on-DM risk model and a Gemini-powered assistant to scale diabetes screening across the national system. Google said the workflow cuts per-case processing from 20 minutes to 25 seconds and will reach a government app used by 10 million people.
Google says Cinematic Video Overviews are rolling out to NotebookLM Ultra users in English. The company says the feature combines Gemini 3, Nano Banana Pro, and Veo 3 to generate more immersive videos than the earlier narrated-slide format.
Google AI Developers has released Android Bench, an official leaderboard for LLMs on Android development tasks. In the first results, Gemini 3.1 Pro ranks first, and Google is also publishing the benchmark, dataset, and test harness.
Google’s February Gemini update packages Gemini 3.1 Pro, Deep Think, Nano Banana 2, Veo templates, and new Canvas tools into one release. The drop shows Google pushing the Gemini app as a front end for reasoning, image, music, and video workflows rather than a plain chat surface.
Google DeepMind said on March 3, 2026 that Gemini 3.1 Flash-Lite delivers faster performance at a lower price than Gemini 2.5 Flash. Google is rolling the model out in preview via Google AI Studio and Vertex AI for high-volume, latency-sensitive workloads.