Google has released new additions to its Gemini Flash model series, targeting developers and enterprise customers building AI agents that require high efficiency, reduced latency, and reliability in production. The company is launching Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber, each tailored for specific use cases such as coding, multimodal tasks, and cybersecurity, respectively. 3.6 Flash and 3.5 Flash-Lite are available immediately through the Gemini API and Gemini Enterprise, while 3.5 Flash Cyber will be accessible to governments and trusted partners via CodeMender in an upcoming limited pilot.
Gemini 3.6 Flash demonstrates a 17% reduction in token usage compared to its predecessor, 3.5 Flash, and achieves lower costs per output token. It shows improved precision in coding, knowledge work, and multimodal tasks, including document parsing and data analysis. Safety safeguards are strengthened, especially in domains like CBRN and cyber offense, reducing risks of misuse. Gemini 3.5 Flash-Lite offers the fastest output in the series at 350 tokens per second and is priced for high-volume, cost-conscious workflows. It outperforms earlier Flash-Lite models in coding and agentic tasks. Gemini 3.5 Flash Cyber is fine-tuned for vulnerability detection and patching, and its deployment is intentionally restricted to prevent misuse.
Google continues to iterate on the Gemini model family, with 3.5 Pro in partner testing and Gemini 4 in pre-training, reflecting its ongoing focus on AI agent infrastructure and safety.