Jul, 134 min to read
Frontend
News · Google puts three experimental Gemini 2.0 models into the Gemini app
Feb, 54 min to read
FrontendGoogle puts three experimental Gemini 2.0 models into the Gemini app
Flash Thinking Experimental, a connected-apps variant, and 2.0 Pro Experimental land in the web and mobile app with different access tiers.
What actually shipped in the app
The announcement bundles three distinct experimental models into one rollout. First, 2.0 Flash Thinking Experimental becomes available to all Gemini app users at no cost. Google describes it as built on 2.0 Flash and trained to break prompts into a series of steps, and claims it is currently ranked as the world's best model.
Second, there's a version of 2.0 Flash Thinking that can interact with YouTube, Search, and Google Maps. Third, Gemini Advanced subscribers get an experimental 2.0 Pro, positioned for complex tasks with better factuality and stronger coding and math performance. All three are rolling out to the Gemini web and mobile app now, with Workspace Business and Enterprise availability described only as coming 'soon.'
The reasoning model shows its work — in the interface
The most concrete frontend detail here is that 2.0 Flash Thinking Experimental displays its thought process. Google's framing is specific: users can see why the model responded a certain way, what its assumptions were, and trace its line of reasoning.
That is a product decision, not just a model capability. Exposing intermediate steps to a general consumer audience means the app now has to render a reasoning trace as part of the ordinary response surface. For a free-tier product, that changes what users learn to expect from an answer — the justification becomes visible alongside the output rather than hidden behind it.
Reasoning meeting the connected apps
The separate variant that talks to YouTube, Search, and Maps is the more forward-looking piece. Google explicitly frames this as exploring how new reasoning capabilities can combine with the app connections that already exist in Gemini.
The distinction matters: rather than folding reasoning and app access into a single model, Google is rolling them out as separate experimental tracks. That suggests the combination of step-by-step reasoning with live app actions is still being tested rather than treated as a finished feature.
Tiering the same generation across free and paid users
The rollout draws a clear line by access level. Flash Thinking Experimental is free to everyone; 2.0 Pro Experimental is reserved for Gemini Advanced subscribers, described as priority access to Google's most capable models.
The specific implication for anyone building on top of the Gemini app: the frontend now surfaces multiple 2.0 models simultaneously, and which one a user reaches depends on their subscription tier and which experimental variant they select. Interfaces built against Gemini should assume a fragmented model landscape — a free reasoning model, a connected-apps variant, and a paid Pro model — rather than a single default, and everything here carries the 'experimental' label, meaning behavior and availability can shift.
Find this story relevant to you?
Contact us to find a unique solution
Need an AI engineering partner that can actually build?
We help businesses integrate AI, build AI-powered products, automate high-value workflows, and modernize the software systems behind them.
Related reading
More analysis around product delivery, operational AI, and the systems work that makes deployment hold up in reality.