Google releases three new AI models, focusing on performance and cybersecurity

Google has released three new AI models: the high-performance Gemini 3.6 Flash, the cybersecurity-focused Gemini 3.5 Flash Cyber, and the agent-oriented Gemini 3.5 Flash-Lite. These models aim to deliver better efficiency at lower costs for developers.

Gemini 3.5 Flash Cyber can find and fix software security holes. In tests, it spotted 55 issues, including 10 that other AI models had never found before. It will only be available to government agencies and trusted partners to prevent misuse.

Google is also working on its flagship model, Gemini 3.5 Pro, which is still in testing. Meanwhile, the company has started pre-training the next-generation Gemini 4, as it faces growing competition from Anthropic and OpenAI.