GOOGLE / GEMINI
THE SHORT VERSION

Google describes stronger agent and coding performance; practical results still depend on the workflow.

What happened

The company announced 3.5 Flash in the Gemini app, AI Mode in Search, Antigravity and the Gemini API. It emphasizes tasks that require multiple steps and tool use.

Why it matters

A model embedded in both search and development tools reaches very different users. The relevant improvement may be completing a task with fewer corrections, rather than producing a more elaborate answer.

The bigger picture

Our take. Evaluate the whole sequence: understanding the request, selecting tools, checking results and recovering from mistakes. Launch benchmarks are a starting point for that assessment.

Primary source: Read the original: Google · Gemini 3.5 ↗

Source checked November 18, 2026. Company claims are attributed; paragraphs marked “Our take” are editorial interpretation.

UNDERSTAND THE IDEA BEHIND THE NEWSWhat makes an assistant an agent↗