Google surfaces Gemini 3.5 Flash

Google entered its I/O developer conference on Tuesday with Gemini positioned across cloud tools, Android software and future XR devices. Google Cloud documentation published this week showed Gemini 3 Flash as an available model, and third-party previews ahead of the keynote pointed to Gemini 3.5 Flash appearing in the Google Cloud Console. (docs.cloud.google.com) Google’s own event pages and Android previews also said Gemini would be central to updates spanning phones, Workspace features and glasses running Android XR. WIRED’s live blog said Google planned updates to its Gemini suite and more detail on Android XR smart glasses. The model branding has expanded alongside the product rollout. (blog.google) Google’s official Gemini pages say Gemini 3 is already being pushed into the Gemini app, Search, AI Studio, Vertex AI and Google’s enterprise agent platform. Google Cloud has separately described Gemini 3 Flash as a lower-latency, lower-cost model for high-frequency workflows, while developers can already call it through Google’s SDKs and CLI tools. ### Where did Gemini 3.5 Flash show up? Google Cloud documentation dated this week lists Gemini 3 Flash as part of the company’s enterprise agent platform. That page says the model combines Gemini 3 Pro reasoning with Flash-class latency, efficiency and cost, and it describes new controls including a `thinking_level` parameter that lets developers tune how much internal reasoning the model uses. (docs.cloud.google.com) Social posts cited in pre-I/O briefings said Gemini 3.5 Flash had appeared in the Google Cloud Console before the keynote. Google had not, in the sources reviewed, published a separate official blog post naming “Gemini 3.5 Flash” by Tuesday morning, but its cloud and product materials showed active rollout of the Gemini 3 Flash family across enterprise and developer surfaces. (blog.google) ### What is Google saying about the model itself? Google’s product post for Gemini 3 Flash says the model offers “Pro-grade reasoning” at Flash speed and lower cost. Google says the model is intended for coding, complex analysis and interactive applications, and that it has become the default model in the Gemini app and AI Mode in Search. (docs.cloud.google.com) Google Cloud’s enterprise post says Gemini 3 Flash is built for near real-time information processing, workflow automation and agentic applications. A separate developer post says the model is available through Google AI Studio and the GenAI SDK, giving developers a direct path from testing prompts to production code. (docs.cloud.google.com) ### How much of this is about multimodal AI? Google has described Gemini as multimodal since the line’s original launch, saying it was built to work across text, images, audio, video and code. In 2024, Google said Gemini 1.5 Pro could support up to a 1 million-token context window in private preview, later extending that long-context push across the Gemini family. (blog.google) Third-party pre-I/O commentary went further, describing Gemini as a natively multimodal system using a mixture-of-experts design and very large context windows. Those descriptions were circulating ahead of the keynote, but the official Google sources reviewed for this story more clearly documented the model family’s multimodal inputs and long-context capabilities than the specific “3.5 Flash” label. (cloud.google.com) ### How far is Gemini spreading inside Google products? Google’s Android preview last week said I/O would include “Gemini Intelligence,” platform upgrades and a glasses preview scheduled for later this year. Google has also said Android XR devices, including glasses and headsets, are being built around Gemini, which the company says can use what the wearer sees and hears for context. (blog.google) The Verge, in a pre-I/O column, said Gemini branding was appearing across Docs, Workspace and other Google products, raising the risk that the name could become an umbrella label attached to too many features. Google’s own Workspace blog has already been rolling Gemini deeper into Gmail, Docs, Meet, Vids and workflow tools. (wired.com) ### What should readers watch for next? Google’s I/O keynote starts on May 19, and the company has directed users to its official I/O and Android channels for additional Gemini and XR announcements. Google’s Android preview said glasses would launch later this year, while WIRED’s live coverage said the keynote would include more details on Android XR smart glasses and Gemini updates. (theverge.com) (blog.google)