In the chatbot market, ChatGPT’s initial dominance is giving way to a more competitive dynamic. Web traffic data released by Similarweb shows that, in twelve months, ChatGPT’s share fell from 87.2% to 68%, while Gemini grew from 5.4% to 18.2%. A clear change, which indicates a progressive redistribution of use between multiple platforms.
The situation, at least for the trend described, reflects the thoughts that many had in the first months after the birth of “commercial LLMs”, when Google seemed to be floundering compared to OpenAI, so much so that CEO Sundar Pichai activated a “red code” at the end of 2022 to catch up on ChatGPT, i.e. when the competing chatbot was only a month old.
Google just needed to come up with a compelling product; the rest would be done by the enormous diffusion of its services.
The growth of Gemini reported by Similarweb appears closely linked to its native integration into the Google ecosystem. Direct presence within Search, Chrome, Android and Workspace reduces friction between user need and access to AI.
The assistant does not in fact require a change of context, but it fits into already consolidated daily flows, encouraging repeated and less episodic use. Think for example of use in Drive, but also more simply of activation with “Hey, Google” on Android phones.
The comparison with Copilot reinforces this reading. Despite integration into Windows and Edge, Copilot remains stuck at around 1.2% share, also equivalent to a drop of 0.3 percentage points compared to 12 months ago.
The data therefore suggests that pre-installation does not automatically guarantee continued adoption. The difference seems to lie in the ability to transform visibility into habit.
Overall, the numbers indicate a stabilization phase of the market. ChatGPT remains the leader, but no longer unchallenged, while Gemini consolidates a relevant position thanks to distribution and frequency of use.
The picture that seems to emerge is that of a sector in which daily accessibility weighs as much, if not more, than the perception of technical superiority. However, especially with the latest releases of the Gemini 3 family and the implementation of the Nano Banana Pro image and Veo 3.1 video model, Google now has products that are excellent in every way.
To conclude everything, now it is OpenAI that has launched a code red, it was talked about a few days before the launch of GPT-5.2, among other things an incremental version of the model which seemed like a hasty response to ward off Gemini’s blows, followed by the launch of the new image model gpt-image-2 to respond to Nano Banana Pro.
Gemini 3 Flash completes the family of Google’s new AI models, fast but incredibly capable
Go to further information
Related News :