Google Search gets a major AI upgrade with Gemini 3.5 Flash
▼ Summary
– Google announced Gemini 3.5 Flash, its latest AI model, at Google I/O, powering AI Mode in Search and the Gemini app for all users.
– For developers, Gemini 3.5 Flash is live in Google Antigravity, the Gemini API, Android Studio, and for enterprise users in the Enterprise Agent Platform.
– Google DeepMind’s CTO stated that Gemini 3.5 Flash rivals large flagship models in intelligence, is four times faster than other frontier models, and outperforms Gemini 3.1 Pro on coding and agentic benchmarks.
– The model delivers frontier-level intelligence at exceptional speed, eliminating the need to trade quality for latency.
– Marketers should monitor how AI Mode responses differ from the previous model for queries important to their sites, as Search is changing rapidly.
At Google I/O today, the company unveiled Gemini 3.5 Flash, its most advanced AI model to date, marking a significant leap forward for Google Search and the broader Google ecosystem. Liz Reid, Google’s head of Search, confirmed that Gemini 3.5 Flash is now the engine behind AI Mode globally, describing it as Google’s “newest Flash model delivering sustained frontier performance for agents and coding.” The model is also being integrated into the Gemini app, and critically, it is available to all users , not just those on a paid tier.
For developers, the rollout is equally expansive. Gemini 3.5 Flash is live in Google Antigravity, the Gemini API for Google AI Studio, Android Studio, and for enterprise users via the Enterprise Agent Platform and Gemini Enterprise. This broad deployment signals Google’s commitment to embedding AI-powered search and coding capabilities across its product stack.
Koray Kavukcuoglu, CTO of Google DeepMind and Chief AI Architect, highlighted the model’s breakthrough performance. “Gemini 3.5 Flash delivers intelligence that rivals large flagship models on multiple dimensions, at the speeds you have come to expect from the Flash series,” he said. He noted that it is “our strongest agentic and coding model yet,” outperforming Gemini 3.1 Pro on demanding benchmarks like Terminal-Bench 2.1 (76.2%), GDPval-AA (1656 Elo), and MCP Atlas (83.6%). It also leads in multimodal understanding with a score of 84.2% on CharXiv Reasoning.
Speed is another differentiator. According to Kavukcuoglu, Gemini 3.5 Flash delivers output tokens at a rate 4 times faster than other frontier models. On the Artificial Analysis index, it lands in the top-right quadrant, proving that frontier-level intelligence no longer requires a trade-off with latency.
Why this matters for you. Gemini 3.5 is already powering AI Mode in Google Search and is poised to soon influence AI Overviews. This represents a clear upgrade over the previous model, and it will continue to evolve in intelligence and utility. As a result, you need to closely monitor how AI Mode responses differ from earlier versions for the queries and prompts that are most relevant to your site. Search is changing rapidly, and staying ahead of these shifts is essential for maintaining visibility and relevance.
(Source: Search Engine Land)
