Google has announced Gemini 3.6 Flash, an updated version of its AI model, following the earlier Gemini 3.5 Flash. The new iteration boasts improved coding capabilities, scoring 49 percent on the Deep SWE test compared to 3.5 Flash's 37 percent. It also shows a modest increase in computer use proficiency, reaching 83 percent on the OSWorld test. A key focus for Gemini 3.6 Flash is efficiency, with Google claiming it uses approximately 17 percent fewer tokens and offers a lower API cost.

In addition to Gemini 3.6 Flash, Google introduced its first AI model specifically for cybersecurity. The company is also not abandoning the Gemini 3.5 branch, releasing Gemini 3.5 Flash Lite and Gemini 3.5 Flash Cyber. Gemini 3.5 Flash Lite is highlighted as Google's most efficient modern AI, capable of processing 350 tokens per second, making it suitable for scaling agentic systems cost-effectively. While benchmarked as nearly on par with frontier models from a year ago, it comes at a competitive price point.