Yesterday was a day of important ads for Google, and apart from deploying Deep Research with Gemini 2.5 Pro in Spain and show its progress on Google Cloud with VERTEX AI, it also gave all the details of Gemini 2.5 Flashits most efficient low cost model that is even faster than Gemini 2.5 Pro.
In January of this year, Gemini 2.0 Flash already deployed as the low -cost AI model that is present in mobile devices for certain AI functions that move locally and protects the privacy of user data, compared to other functionalities that do require the use of mobile data to access servers where they are processed.
The faster and more efficient an AI model, the easier it will be than Everything is processed on the mobileand so the fear that a large part of users are suffering when they make more precise consultations that give more information about their daily lives and chores are eradicating.
Gemini 2.5 Pro has surprised its own and strangers by surpassing Chatgpt and Claude in certain tasks such as the programming and creative writingand for a few weeks many users are going to Gemini Advanced to use it. That Gemini Flash 2.5 uses Gemini 2.5 Pro is great news, because it becomes a great performance model with that special point of lower use of resources.
Google did not want to give data on the number of specific parameters of the Flash version, but 2.5 Pro (experimental) has been the first Gemini model to implement the dynamic thinkinga technique that allows the model to modulate the amount of simulated reasoning to an answer. And the important thing: 2.5 Flash is also a reasoning model, but a little more advanced.
Deep Research Benchmarks enhanced by Gemini 2.5 Pro
The free android
The other important gemini 2.5 flash point is its Flexibility with granular controlespecially for developers and third parties, since it allows to refine the speed, precision and cost balance for the specific needs of the developer in its project.
Gemini 2.5 Flash will not land at the moment in the Gemini app that we have on our mobiles or web version, and He stays on the VERTEX AI developer platform. Here Google will be able to collect the data and how users have interacted with the new model, which will allow them to properly develop and optimize Gemini 2.5 Flash before it reaches the mobile app.
The free android
An important day for Google, not only for the deployment of Deep Research for research through 2.5 Pro in Spain, but by flash 2.5 announced from Your blogsince it is the model that we will supposedly see in the mobiles and that will promote new experiences that can be produced through Gemini or Samsung’s very Galaxy ai. Everything is going very fast and the Google machine seems to start being greased in front of OpenAi with its chatgpt.
