Google I/O: ‘Spark’ Agent and ‘Gemini 3.5 Flash’ Model Unveiled Google used its 2026 I/O keynote to push a future of “always‑on” AI agents, but the event was shaped as much by what didn’t launch as by what did.

Early briefings: Spark, the 24/7 agent

Ahead of the keynote, reporters were told Google was rolling out its first truly “always-on” AI agent, Spark, able to “plan parties, collate notes, and run a variety of other tasks in the background” while running 24/7 on Google Cloud rather than a user’s laptop. Positioned as a shift from chatbot to proactive assistant, Spark can read from Gmail and Docs and will later connect to Chrome and third‑party tools, letting it draft emails or track RSVPs autonomously.

Onstage: Flash arrives, Pro delayed

When the keynote began, the first major model news was not the long‑rumored flagship. Instead, Google unveiled Gemini 3.5 Flash, described as its strongest model yet for coding and autonomous agents, able to execute pipelines and even build an operating system from scratch in internal tests. DeepMind’s Koray Kavukcuoglu said Flash “outperforms our latest frontier model, 3.1 Pro, on nearly all the benchmarks” and is up to 12x faster in an optimized configuration.

Developers evangelized the speed: “Meet Gemini 3.5 Flash! ⚡️ It’s fast, great for building rich UIs + agents and is stronger from coding to long-horizon tasks and multi-step workflows,” wrote Google engineer Addy Osmani on X. CEO Sundar Pichai followed, highlighting that Flash is “available today” across Google products and APIs and “better across almost all benchmarks with huge progress in coding.”

But anticipation around Gemini 3.5 Pro turned into audible disappointment when Pichai told the I/O crowd the more powerful model “isn’t coming out yet” and asked developers to “give us until next month,” despite saying it was already “showing great improvements.”

Ecosystem and competition reactions

Coverage framed Spark as Google’s most ambitious agentic push yet, tightly integrated with Workspace and running on dedicated cloud VMs, essentially acting like a human colleague reachable via a dedicated Gmail address. Another analysis argued Gemini 3.5 Flash could finally make “complex agentic tasks worth doing at scale” by combining near‑frontier quality with much lower cost and latency.

Yet competitors and skeptics quickly surfaced. A TechCrunch piece said Google is “betting its next AI wave on agents, not chatbots,” emphasizing multi‑hour autonomous runs and banks automating multi‑week workflows with Flash. Meanwhile, a benchmark thread on X—boosted by Elon Musk with the comment “Try Composer 2.5”—noted Gemini 3.5 Flash ranked only tenth on one coding benchmark, below some rivals.

Together, the announcements and reactions underscore Google’s pivot: Gemini 3.5 Flash and Spark are here now to power agents, while the more powerful Gemini 3.5 Pro, and the broader race with competing models, still loom over what comes next.

Continue reading https://foxvector.com