Google DeepMind has introduced three new Gemini Flash models focused on reducing costs and improving speed for AI agents, while the high-end Gemini 3.5 Pro remains in partner testing since February.
The Gemini 3.6 Flash model, succeeding Gemini 3.5 Flash, delivers up to 17% fewer output tokens, enhancing efficiency at a price of $1.50 per million input tokens and $7.50 per million output tokens. This model excels in coding, reasoning, and document analysis and is already used by clients such as Hebbia and Harvey for parsing documents and drafting reports.
For tasks that require high volume and fast processing, the Gemini 3.5 Flash-Lite offers the lowest cost at $0.30 per million input tokens and $2.50 per million output tokens, running at 350 output tokens per second. Both 3.6 Flash and Flash-Lite are currently available on Google’s developer platforms.
The third model, Gemini 3.5 Flash Cyber, targets cybersecurity applications by scanning code to detect vulnerabilities using the CodeMender tool. In tests on the V8 JavaScript engine, Flash Cyber identified 55 confirmed vulnerabilities compared to 47 by Gemini 3.5 Flash and 36 by Anthropic’s Claude Opus 4.6. Due to safety restrictions, Flash Cyber is limited to governments and trusted partners and is not publicly accessible.
