Google launches Gemini 3.5 with advanced agentic AI capabilities
Agentic AI systems are becoming central to Google’s long-term AI strategy.
Google has announced Gemini 3.5, a new family of AI models designed to combine frontier-level reasoning with stronger agentic capabilities across consumer, developer and enterprise products.
The company is launching the series with Gemini 3.5 Flash, which it describes as its strongest agentic and coding model so far. Google said the model is built for complex, long-horizon tasks, including multi-step workflows, coding, document analysis and enterprise automation.
According to Google, Gemini 3.5 Flash outperforms Gemini 3.1 Pro on several coding and agentic benchmarks, including Terminal-Bench 2.1, GDPval-AA and MCP Atlas, while also improving multimodal reasoning. The company also claimed the model is significantly faster than other frontier models when measured by output tokens per second.
Google said the model can support agentic workflows through its Antigravity development platform, including the use of collaborative subagents for coding, financial document preparation, data analysis and other complex enterprise tasks. It also highlighted richer multimodal capabilities, including interactive web interfaces, visual reasoning and generative UI experiences.
Gemini 3.5 Flash is available through the Gemini app, AI Mode in Google Search, the Gemini API in Google AI Studio, Android Studio, Google Antigravity and Gemini Enterprise products. Google also said Gemini 3.5 Pro is being used internally and is expected to roll out next month.
The announcement also introduced Gemini Spark, a personal AI agent powered by Gemini 3.5 Flash. Google said Spark is designed to run continuously, helping users navigate digital tasks and take action under their direction. It is starting with trusted testers, with a beta planned for Google AI Ultra subscribers in the US.
Alongside performance improvements, Google said Gemini 3.5 was developed under its Frontier Safety Framework. The company said it strengthened safeguards related to cybersecurity and chemical, biological, radiological and nuclear risks, while adding safety training and interpretability tools intended to reduce harmful outputs and mistaken refusals.
The launch reflects a wider industry shift from conversational AI assistants towards systems that can plan, coordinate and execute tasks across digital environments, with major AI companies increasingly competing on agentic workflows, coding performance, multimodal interaction and enterprise integration.
Why does it matter?
Gemini 3.5 shows how the AI race is moving beyond chatbot performance towards systems that can act across software, workflows and enterprise environments. Faster agentic models could help automate coding, analysis and business operations, but they also raise governance questions around supervision, safety, accountability and how much autonomy users and organisations should give to AI agents.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!