Google is expanding its family of AI models by releasing three new versions of Gemini Flash at once: Gemini 3.6 Flash, 3.5 Flash-Lite, and the specialized Gemini 3.5 Flash Cyber. As a market analyst, I see a clear signal in this move: the company is betting on efficiency, accessibility, and security, which is critically important for AI adoption in the corporate sector.

Gemini 3.6 Flash: Lower Costs, Greater Accuracy

The new flagship model of the lineup, Gemini 3.6 Flash, is positioned as the primary workhorse for coding, data handling, and multimodal tasks. The key advantage is a 17% reduction in output token consumption compared to 3.5 Flash, and in DeepSWE tests, the savings reach an impressive 65%. The cost is $1.50 per 1 million input tokens and $7.50 per 1 million output tokens, making it cheaper than its predecessor.

Improvements in benchmarks confirm performance growth:

  • DeepSWE: 49% vs. 37% for 3.5 Flash;
  • MLE Bench: 63.9% vs. 49.7%;
  • OSWorld-Verified: 83.0% vs. 78.4%.

These metrics directly enhance computer use capabilities, which are now available as a built-in client-side tool via the API and the Gemini Enterprise application.

Gemini 3.5 Flash-Lite: Speed and Cost Efficiency for Agents

The 3.5 Flash-Lite model, according to Google, is the fastest and cheapest in the lineup. Its speed reaches 350 output tokens per second at a price of $0.3 per 1 million input tokens and $2.5 per 1 million output tokens. In agentic scenarios, it significantly outperforms 3.1 Flash-Lite, and in tests like SWE-Bench Pro (54.2% vs. 49.6%) and OSWorld-Verified (74.0% vs. 65.1%), it even surpasses 3 Flash. This makes it an ideal choice for tasks where low latency and minimal costs are critical.

Gemini 3.5 Flash Cyber: AI Security Specialist

Special attention deserves Gemini 3.5 Flash Cyber. This model, based on 3.5 Flash, is aimed at finding and fixing cyber vulnerabilities. In CodeMender, it functions as a group of agents that generate a unified report. Its performance on the CyberGym benchmark is competitive, but due to its dual-use nature, access at launch is restricted: only governments and trusted partners within a pilot program.

Availability and Future Plans

Gemini 3.6 Flash and 3.5 Flash-Lite are already available to developers via the Gemini API in Google AI Studio and Android Studio. For businesses, the models are deployed in the Gemini Enterprise Agent Platform, and 3.6 Flash is in Gemini Enterprise. Regular users can try them through the Gemini app, and 3.5 Flash-Lite is already being integrated into Google Search. Google also confirmed that the flagship Gemini 3.5 Pro is being tested by partners, and Gemini 4 is in the pre-training phase.

My analysis: The release of Gemini 3.5 Flash Cyber is a strategic move by Google that could fundamentally change the cybersecurity landscape. If the model proves its effectiveness in real-world conditions, we will see mass adoption of AI agents for automating vulnerability discovery, which will reduce the burden on security teams and shorten response times to threats. Restricting access at launch is a sensible measure, but in the long term, I expect such tools to become the standard for all major organizations.