Google launches three new Gemini AI models with lower cost and enhanced security

Google announced three new Gemini AI models – Gemini 3.6 Flash, Gemini 3.5 Flash‑Cyber and Gemini 3.5 Flash‑Lite – expanding its generative‑AI portfolio. The flagship 3.6 Flash reduces token usage by about 17 % compared with earlier versions and is priced at roughly $1.50 per million input tokens and $7.50 per million output tokens (or equivalent euros). The model shows higher accuracy on coding benchmarks (up to 65 % improvement on DeepSWE) and better performance on tasks such as image, video and PDF handling.

Gemini 3.5 Flash‑Cyber is tailored for vulnerability detection; in internal tests it identified 55 security flaws in the V8 JavaScript engine, including ten critical issues not previously spotted. Access to this model will initially be limited to governments and trusted partners to allow defensive teams to patch flaws before malicious actors can exploit them.

Gemini 3.5 Flash‑Lite is a lightweight version designed for high‑throughput digital‑assistant and autonomous‑agent workloads, offering the fastest token‑per‑second rate in the Gemini family.

Separately, users reported a brief service disruption on 24 July 2026, with many experiencing connection time‑outs and slow responses. No official statement has clarified whether the outage was global or regional, but it appears to have been a temporary infrastructure issue.