Google launches three new models of its AI and is already laying the groundwork for Gemini 4

Google has taken another step in the arms race for control of the AI market with the launch of its new Gemini models, which promise to be more powerful and more cost-effective than their predecessors. These are three new models with very different goals: one general-purpose model, another focused on efficiency, and a third specialized in cybersecurity.

Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber arrive

It is clear that Google is not going to be the one to solve the very serious naming problem we have across the AI market, with three models belonging to two different Gemini generations, which certainly does not make its catalog any easier to understand.

On one hand, there is 3.6 Flash, which now becomes Google’s main model. According to the company, 3.6 Flash “offers better performance in coding, knowledge work, and multimodality” than its predecessor. In addition, it reduces output token usage by 17% compared with 3.5 Flash.

According to Google’s documentation, Gemini 3.6 Flash is useful for users, developers, and businesses alike. The use cases mentioned include agent-based workflows, programming tasks, and business processes lasting several weeks. In any case, Google itself warns that the new model is far from infallible, and it may display some of the general limitations of foundation models, such as hallucinations.

The other two new releases are Gemini 3.5 models:

• 3.5 Flash-Lite: Google describes it as its “fastest and most cost-effective” model in the 3.5 class, outperforming earlier generations “significantly” in agent-based workflows. Those who have used 3.5 Flash-Lite highlight its combination of speed, intelligence, and cost-effectiveness for scaling agent-based workflows and data processing tasks
• 3.5 Flash Cyber: This is a new cybersecurity-focused model, optimized to identify and fix cybersecurity vulnerabilities at a lower per-token cost than larger models.

What about Gemini 3.5 Pro?

Although the market had been expecting the launch of Gemini 3.5 Pro, its arrival will take a little longer. According to Google, the model is currently in testing with partners and “plans to release it to the general public as soon as it is ready.” In any case, Google has explained that it has already begun the pretraining phase for the new Gemini 4 generation. “We are very pleased with the results,” the company said.

For now, these new releases will have to do. The 3.6 Flash and 3.5 Flash-Lite models are already available starting today in Gemini.

Image: ChatGPT