Pixel 9 Pro XL review: Android with AI running locally, Gemini Advanced, and the best photography

The new generation of mobile phones Pixel of Google It has arrived this year a little earlier than usual, and that prevents them from doing so accompanied by the next version of Androidthe 15th, which will arrive this fall. But that doesn’t mean that they are short of new features. There are some in terms of software, the integration of artificial intelligence with Android running on the device itself, the design, the hardware and its renowned photographic capacity, revised both from the point of view of the cameras and the software. At La Razón we have been able to test the largest of the family, the Pixel Pro 9 XLand these are our impressions.

New design, more elegant and premium

The first thing that catches your attention in the new Pixel 9 is changes in design. The metal frame is now flat, which improves grip and gives it a more elegant and robust appearance. The rear bar that houses the camera modules is reduced in size and no longer goes from edge to edge, and maintains the dimensions of the Pixel 8 Pro with a slightly larger screen, 6.8 incheswith thinner bezels. This is a very subjective aspect, but together with a matte glass back, they make it in the prettiest Pixel that Google has released and the most premium looking oneIt is a bit heavy, 221 grams, but can be handled comfortably even with one hand on most occasions, despite the screen size.

This is a LTPO OLED with a resolution of 1,344 x 2,292 pixels and a refresh rate of 120 Hz. It is like that of the Pixel 8 Probut slightly larger and with greater luminosity. It reaches 3,000 nits compared to 2,400 of the previous generationmaking its outdoor use one of the best offered by a smartphone today.

Inside, the SoC G4 Tensioner It succeeds the G3 and, according to the performance tests that are becoming known, it only offers a slight improvement over the previous generation, around 10%. It’s not something the average user is going to notice as a shortcoming or miss more power. The phone performs smoothly at all times, and also in the AI-related tasks we’ll get to later. That’s helped by its generous 16 GB RAMone of the most necessary requirements for AI models to be able to run locally on mobile phones. Now, it is a mobile phone with 7 years of guaranteed operating system and security updates from Google, and a very reasonable doubt is whether this hardware will still be sufficient for the Android that the company launches in 2030 or 2031.

Gemini Nano and AI running locally on the device

With this section we come to one of the cruxes of the matter. Gemini Nano is Google’s version of generative AI that is lighter and can run locally on devices. This approach is interesting because it allows users’ privacy to be protected much better than with models that run in the cloud, and what Google has done is integrate it with Android more fully than it did on the Pixel 8.

If in the previous generation, Nano only worked on the keyboard Gboard and with WhatsAppnow does so with more sections of the mobile. For example, Notes from the call is a feature that generates a summary and a complete transcript of the conversation that has been had on the phone. Calls that also have a feature of automatic subtitles processed locally, just like the previous one. Another new task is to generate the smart responses in the app Messages from Google, something that was previously done from the cloud. Pixel Screenshots is a new Pixel-exclusive app that uses Gemini Nano to process your image library and generate insights to help you perform natural language searches and easily find what you want.

Gemini Nano, which is multimodal and can be interacted with through images, text, or voice, is not the only AI running locally on the Pixel 9. The app Pixel Studioalso exclusive to the Pixels, allows you to create images from text using either a template that is downloaded and run locally on the device or by connecting to the cloud to use Image 3Google’s latest text-to-image language model.

Gemini Advanced

Nano is the smallest language model managed by Google under the Gemini brand umbrella. Gemini AdvancedGoogle’s most advanced model, is available on the terminal through the promotion that the Pixel 9 brings, consisting of a one-year free trial Google One AI Premium. This Google subscription plan, which costs €21.99 per month, gives you access to a series of benefits such as more storage space, longer video calls on Google Meet, or the latest Google AI models, among others.

If you use the promotion, you not only have access to the Gemini model in the cloud through the Google Assistant, which can be configured to be either traditional or Gemini, and its dedicated app, but it also allows you to use AI with other Google apps like Gmail, YouTube, Maps, or Docsamong others, and more to come in the future. Is it very different from the usual user experience with any of them? Yes, by a long shot. An assistant that you can ask to find emails, summarize them, and write replies greatly streamlines email management. One may have doubts about whether it is reliable enough to delegate this type of task, given that generative AIs are prone to fabrication and providing misinformation. Google makes the usual warnings about the mistakes their AI can make, but in my experience they have better results when working with a closed and reduced set of data, such as emails stored in Gmail, than when they have to interpret what they find on the Internet.

Other examples: its integration with YouTube allows you to ask it to find videos on a topic, give you a selection, ask it to summarize them and, if you are interested, proceed to watch them. It is another quick and easy way to save time on platforms that are prone to waste time. You can also tell it that you want to go on a trip to a certain place and have it find the route for you on Maps. Gemini will show you all the necessary information, load the route into Maps and show it to you integrated into its response and with a link to Maps. Documentsis a writing assistant that helps you create any content.

The way to access the assistant is either through its dedicated app or by holding down the power button. Following this last method, Gemini overlays anything you have on screen and among its options will be available to ask about what you are viewing at that moment. Depending on whether you are subscribed to Google One AI Premium or not, Gemini or Gemini Advanced will answer you. The first, more limited, is not, for example, connected to the aforementioned applications.

Does all this significantly improve the user experience? Yes, and AI applications can significantly improve the user experience. It is the most significant advance that mobile phones have had in many yearsThere is still a long way to go, but AI is putting them on the path to being more intuitive and more capable.

Photography: Software rewritten from scratch

More than anything else, the photography section is what has earned the Pixel brand one of the best cameras on the market, but this is not entirely accurate. In addition to the quality of the camera itself, Pixel has been standing out for several years for its software. A series of tools that treat the photographs taken as the raw material on which a series of software modifications are applied to achieve a specific image.

An important aspect is that, while the Pixel 9, the entire range, maintains a very similar set of cameras to that of the 8 Pro (50 MP main, 48 MP wide angle and 48 MP telephoto that the Pixel 9 loses), The camera software has been rewritten from scratchAmong the new features that have been added, there is Include mewhich allows you to take a picture, pass your phone to another person and join the frame so that everyone present is included in the image. Magic Editor Google Photos now has a feature Reimagine which allows you to select a part of the image and fill it with AI-generated content. And Automatic Framing automatically improves the composition of a photo, among other new features.

The Pixel 9 allows you to manipulate an image in a thousand ways, and Software changes allow for sharper resultsaccording to Google. In any case, it is clear that more traditional photography without image alterations by software is on its way to becoming a thing of the past.

The Pixel 9 takes the title of AI phone of the year again and right now They have the lead in terms of AI features and integration with the operating system and apps.It remains to be seen whether and how Apple Intelligence iPhones will make it to Europe, but for now there’s no denying that the Pixel 9 is The best AI phone available on the market.