Google Tests 4 Gemini 3 Pro Variants, Expands AI to Maps

The testing, as reported by Mastodon Social ML Timeline, suggests Google is exploring multiple specialized versions of the model, likely designed for different tasks or computational demands. This multi-pronged approach indicates a strategy to deploy Gemini 3 Pro across its vast ecosystem of products, tailoring its capabilities for specific applications rather than relying on a single, monolithic AI.

This expansion is already underway with the launch of the new ‘Ask Maps’ feature, which integrates Gemini’s conversational AI directly into Google Maps. This move represents one of the most significant integrations of generative AI into a core Google product used by billions, transforming the service from a static directory into an interactive, query-based guide.

The aggressive push into AI comes as Google faces mounting regulatory pressure in Europe. According to Reuters, the company was recently hit with a massive €3.45 billion antitrust fine from the EU concerning its adtech practices. In a separate but related development, Reuters also exclusively reported that Google’s new AI Overviews feature is the subject of an antitrust complaint from a consortium of German publishers, represented by VG Media. The publishers allege they cannot opt out of having their content used to train and fuel Google’s AI models without also disappearing from search results entirely, a predicament that threatens their web traffic.

Amidst this regulatory scrutiny, the broader AI industry is displaying unexpected collaboration. In a rare move, rivals including Google, OpenAI, and Anthropic are partnering on a new startup accelerator based in Paris named F/ai, as reported by Mastodon Social ML Timeline. This joint effort to foster the next generation of AI ventures signals a strategic, if cautious, shift towards cooperation on foundational issues even while the companies compete fiercely in the consumer market.

That competition is increasingly focused on monetization. According to The Register, while Google is not currently showing ads within its standalone Gemini chatbot, it is exploring how to integrate advertising into its broader “AI Mode.” This places Google in a similar race as its rivals OpenAI and Anthropic, who are also wrestling with how to effectively generate revenue from their expensive-to-run AI technologies. The key difference, as noted by The Register, is Google’s immense advantage in existing advertising infrastructure; the company already knows how to make money from information retrieval at a scale others can only imagine.

The simultaneous testing of multiple Gemini variants, the expansion into Maps, and the experimentation with ad models paint a picture of a company in a furious sprint. Google is leveraging its unparalleled integration across search, mobile, and advertising to embed AI into the very fabric of its services, all while navigating a complex web of regulatory challenges and an unlikely detente with its fiercest competitors. The goal is clear: to ensure that as the world’s interaction with technology becomes increasingly conversational, the answers still come from Google.