Google has introduced three new Gemini AI models focused on delivering faster performance and lower operating costs, as the company shifts its strategy toward efficiency rather than simply building larger, more powerful models.
The announcement includes Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber, all designed to improve AI agents while reducing computing costs. According to Google, Gemini 3.6 Flash cuts token usage by up to 17% compared with its predecessor, helping businesses lower AI expenses.
Gemini 3.5 Flash-Lite is Google's fastest and most affordable 3.5-series model, while Flash Cyber targets cybersecurity tasks by detecting and fixing software vulnerabilities at a lower cost than competing models. Initially, Flash Cyber will only be available to governments and selected partners.
Despite the new releases, Gemini 3.5 Pro remains unavailable. Originally expected in June, the flagship model is still undergoing testing with partners and will launch only when Google says it is ready.
Google also confirmed it has begun the pre-training process for Gemini 4, its next-generation flagship AI model, although its release is still months away.
The company says growing concerns over AI spending are driving demand for more efficient models, positioning its latest Gemini lineup as a balance between performance and cost.