From Gemini 3.5 Flash to $2 million hackathon, everything Google announced at I/O 2026

Google on Tuesday unveiled a raft of artificial intelligence products and developer tools at its annual I/O 2026 conference, led by the launch of Gemini 3.5 Flash, a faster and more capable AI model that would power autonomous software agents and real-world coding applications.

The company also introduced a revamped AI development platform called Antigravity 2.0, managed agents in the Gemini API, native Android app development support in Google AI Studio, and a global hackathon with a $2 million prize pool.

Google Chief Executive Officer Sundar Pichai said the announcements marked the beginning of the "agentic Gemini era", where AI systems move beyond responding to prompts and are able to independently plan, reason and perform tasks.

“We’re now in the part of the AI cycle where people want to see the value in the products they use every day,” Pichai said. “As we enter this agentic era, Search will be more helpful and powerful than ever.”

Gemini 3.5 Flash

Google said Gemini 3.5 Flash combines "frontier intelligence with action" and outperforms Gemini 3.1 Pro across most benchmarks while running four times faster than rival frontier models.

The company said the model offers frontier-level performance at less than half the price of comparable models and is available immediately across Google products and application programming interfaces (APIs).

“We’ve heard that many companies are already blowing through their annual token budgets, and it’s only May. If companies used a mix of Flash and other frontier models they could save a lot of money,” Pichai said.

Google also said Gemini 3.5 Pro, an enhanced version of its flagship model, will be released next month.

Antigravity 2.0 introduced

Google expanded Antigravity, its AI-first development platform, with the launch of Antigravity 2.0, a standalone desktop application that allows developers to orchestrate multiple AI agents to work in parallel.

Pichai said Antigravity was "expanding beyond the coding environment, turning it into a platform to develop and manage cohorts of autonomous AI agents."

The application includes dynamic sub-agents, scheduled automation tasks and integrations with Google AI Studio, Android and Firebase.

The company also launched Antigravity CLI for terminal users and an Antigravity software development kit (SDK) that lets developers define and host custom agents on their own infrastructure.

For enterprise customers, Antigravity is being integrated with Google Cloud projects through the Gemini Enterprise Agent Platform.

Managed Agents in Gemini API

Google introduced Managed Agents in the Gemini API, enabling developers to create an AI agent with a single API call.

The agents can reason, use tools and execute code in isolated Linux environments with persistent storage, making it possible to run multi-step and multi-session workflows.

Developers can also customise these agents using markdown-based instructions and templates available in Google AI Studio.

Google AI Studio gets Android and mobile support

Google AI Studio, its browser-based development environment, received a series of upgrades.

The company said a new mobile app will be available for pre-registration this week, allowing developers to prototype applications while on the move.

Google also added native Android development support, enabling developers to build Android apps from text prompts and publish them directly to Google Play Console test tracks.

Projects created in AI Studio can now be exported to Antigravity with a single click, while agents will also be able to access Google Workspace APIs such as Docs, Gmail and Sheets.

New subscription plan

Google launched a new Google AI Ultra subscription priced at $100 per month.

The plan offers five times higher usage limits in Antigravity compared with the Google AI Pro plan and includes bonus credits for AI development workloads for a limited period.

Gemini Spark personal AI agent

Among consumer-facing announcements, Google introduced Gemini Spark, a personal AI agent within the Gemini app that can work continuously in the background on dedicated Google Cloud virtual machines.

The agent will integrate with Google's own products and, later, with third-party tools through the Model Context Protocol (MCP).

Spark is being rolled out to trusted testers and will be available in beta next week to Google AI Ultra subscribers in the United States.

Gemini Omni and multimodal AI

Google also announced Gemini Omni Flash, a new model capable of generating outputs across different modalities from a range of inputs, beginning with video.

“This new model combines Gemini’s intelligence with our generative media models — a huge leap forward in world understanding,” Pichai said.

The model will be available in the Gemini app, Google Flow, YouTube Shorts and through APIs in the coming weeks.

Search, YouTube and Workspace updates

Google shared updated usage figures, saying AI Overviews in Search now has more than 2.5 billion monthly active users, while AI Mode has crossed 1 billion users.

The Gemini app has surpassed 900 million monthly active users, up from 400 million a year ago.

New product features announced include Ask YouTube, which allows users to ask questions and jump to relevant moments in videos, and Docs Live, a voice-powered feature that creates and edits documents using spoken instructions.

Deepfake detection partnerships

Google said its SynthID watermarking technology has now marked more than 100 billion AI-generated images and videos.

“And now we’re going a step further and adding Content Credentials verification across products. This will show you if the origin of the content was AI or a camera, and if it’s been edited with generative AI tools,” Pichai said.

The company announced that OpenAI, ElevenLabs and Kakao will adopt SynthID to help identify AI-generated content.

Record infrastructure spending

Google said it is dramatically increasing spending to support the effort, forecasting about $190 billion in capital expenditures this year, up from $31 billion in 2022, largely to fund data centers and custom AI chips known as Tensor Processing Units (TPUs).

The company highlighted its eighth-generation TPUs, including TPU 8t for training and TPU 8i for inference.

$2 million hackathon

Google announced the "Build with Gemini XPRIZE Hackathon", offering a $2 million prize pool, which the company said is the largest ever for a hackathon.

Finalists will present their projects at the Moonshot Gathering in Los Angeles in September.

Google said over 8.5 million developers are now building applications with its AI models every month, while its APIs process around 19 billion tokens per minute.