Google has officially introduced Gemini 3.5, its latest family of AI models focused on combining “frontier intelligence with action.” The announcement was made during Google I/O 2026 as part of the company’s broader push toward agentic AI systems capable of handling complex multi-step tasks autonomously.  

The first model in the lineup, Gemini 3.5 Flash, is available starting today and is becoming the default model across the Gemini app and Google Search AI Mode. Google says the model is optimized for speed, long-horizon reasoning, coding, and AI agent workflows.  

According to Google, Gemini 3.5 represents a major step toward more capable AI agents that can perform actions rather than simply respond to prompts. The company specifically highlighted improvements in:

• Agentic task execution

• AI-assisted coding

• Long-context reasoning

• Interactive UI generation

• Multi-step workflow handling

• Real-world utility tasks

Google described Gemini 3.5 Flash as its “strongest” model yet for agentic behavior and coding performance. The model is designed to balance low latency with stronger reasoning capabilities, making it suitable for applications that require continuous task execution and orchestration.  

The release is closely tied to Google’s newly announced “Agentic Gemini Era,” where AI systems increasingly operate proactively across products like Search, Android, Workspace, YouTube, and Cloud.  

During the I/O keynote, Google also demonstrated Gemini Spark, a new AI agent system powered by Gemini 3.5 Flash. Spark is designed to execute tasks independently without requiring users to keep devices active continuously.  

Another major focus of Gemini 3.5 is coding. Google says the model delivers improved agentic coding capabilities, helping developers generate richer web interfaces, automate workflows, and manage longer development tasks more effectively.  

The launch also includes stronger safety systems. According to Google, Gemini 3.5 Flash introduces improved guardrails that reduce harmful generations while lowering false-positive safety blocks on legitimate prompts.  

Google confirmed that Gemini 3.5 Pro, the company’s higher-end reasoning model, is expected to launch next month. CEO Sundar Pichai acknowledged during I/O that developers were eagerly waiting for the release, but the company decided to delay rollout slightly for further refinement.  

The broader Gemini ecosystem is also expanding rapidly. Alongside Gemini 3.5, Google introduced:

• Gemini Omni multimodal models

• Gemini Spark autonomous agents

• AI-powered Search upgrades

• Gemini Live integrations

• New Android XR experiences

For developers, Gemini 3.5 signals Google’s continued shift from traditional chatbot experiences toward AI systems capable of planning, reasoning, executing tasks, and interacting with software environments more autonomously.

Google says Gemini 3.5 Flash is already rolling out globally across Gemini apps, Google AI Studio, and enterprise AI platforms.  

Developers can learn more through Google’s official resources:

• Google Official Gemini 3.5 Announcement