Google I/O 2026: Gemini 3.5 Flash Arrives with Flagship Performance at Flash Speed

Google launched Gemini 3.5 Flash at its I/O 2026 keynote on May 19, making it generally available the same day through Google AI Studio, Android Studio, and the Antigravity platform.

Outperforming the Flagship

Despite sitting below the forthcoming 3.5 Pro, Gemini 3.5 Flash beats Gemini 3.1 Pro on challenging coding and agentic benchmarks:

• Terminal-Bench 2.1 (coding): 76.2%
• GDPval-AA (real-world agentic tasks): 1,656 Elo
• MCP Atlas (tool-use reliability): 83.6%
• CharXiv Reasoning (multimodal understanding): 84.2%

Compared to Gemini 3.1 Pro, it delivers output 4x faster and costs roughly 40% less -- $1.50 per million input tokens and $9.00 per million output tokens.

New Default for Search AI Mode

Google is rolling Gemini 3.5 Flash out as the default model for Search AI Mode, which now exceeds one billion monthly users globally. The model is tightly integrated with Google's Antigravity agent platform for long-horizon agentic workflows.

DeepMind CEO Demis Hassabis drew attention at the keynote with the remark: "We're at the foothills of the singularity." He argued that the rapid rise of agentic AI systems makes human-level AI by around 2030 a realistic timeline. The comment sparked significant debate across the tech industry.

Gemini 3.5 Pro is slated for release next month. Details at the Google blog.