Google expanded its artificial intelligence lineup this week with the release of three new models: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini Flash Cyber, the latter built specifically for cybersecurity tasks. The rollout fits into the company's ongoing approach of delivering steady, incremental updates across its generative AI ecosystem.
Flash Cyber stands out for its focus on threat detection and vulnerability analysis, an area where Google appears eager to compete with other AI developers already offering similar tools for security teams. Meanwhile, Gemini 3.6 Flash and 3.5 Flash-Lite are geared toward speed and computational efficiency, favoring fast responses over deep reasoning power.
What stood out most from this announcement was a conspicuous gap: Gemini 3.5 Pro, the flagship-tier model of this generation, still hasn't been released. Its continued absence, even as Google confirmed it is already training the next full generation, Gemini 4, raises questions about the company's internal priorities and when that higher-capability Pro version — long expected by developers handling more demanding workloads — will finally arrive.
Google also teased, without offering a specific timeline, that Gemini 4 is currently in training. That detail suggests the company could be moving toward its next generation before ever completing the 3.5 lineup. This isn't officially framed as a roadmap overhaul, but it does align with an increasingly fragmented release pattern, where Google seems to favor lightweight, specialized models over mid-tier, high-capability releases.
For users and developers relying on the Gemini ecosystem, the immediate takeaway is an expansion of practical options — speed, efficiency, and security-focused tools — while the flagship model of the 3.5 series remains on hold. Attention now shifts to Gemini 4, whose training is already underway, though no release date has been shared.