On Tuesday, Google DeepMind unleashed its latest trio of Gemini models: the workhorse 3.6 Flash, the budget-friendly 3.5 Flash-Lite, and the cybersecurity-focused 3.5 Flash Cyber. These updates promise enhanced efficiency in coding and knowledge work while reducing token usage by up to 17%, making them cheaper than their predecessors.

The 3.5 Flash Cyber will be available exclusively to governments and trusted partners as part of a limited access pilot program, according to Google. The company is currently testing the Gemini 3.5 Pro with partners and hopes to launch it soon, although internal delays have been reported.

Despite these updates, the absence of the long-anticipated Gemini Pro model is noteworthy. In this fast-paced AI race, OpenAI has rolled out GPT-5.5 and GPT-5.6, while Anthropic has launched Claude Opus 4.8, Claude Sonnet 5 and expanded access to its Fable 5 model.

Logan Kilpatrick, Google DeepMind’s product lead, mentioned that Gemini Pro is Google’s highest-capability offering for complex reasoning and coding tasks, while the Flash models prioritize lower cost and faster response times for production applications. The company hopes to deliver efficiency, latency, and reliability to customers building AI agents at scale.

The launch of these new Gemini models marks a significant stride in the ongoing AI arms race but also highlights the challenges Google faces in maintaining its competitive edge.