Within the framework of the Mobile World Congress 2025, Xiaomi has announced great plans for 2025. Xiaomi 15 Ultra is one of them and, together with Xiaomi 15make up the new high range for this year. These two phones arrive with Hyperai, the interpretation of the Chinese manufacturer. Xiaomi has come something after his rivals to integrate multiple functions of AI so natively, but says the saying that better.
I have been able to try some of these first -hand functions, so I’m going to tell you what Xiaomi has implemented on these phones, and what to expect from this first stage.
This is how Xiaomi money earns – they attract you and catch you
The novelties of Hyperai
Hyperai’s novelties are divided into seven blocks. In all of them there is a common point: it is a native integration. In other words, they are deep integrations at the system level and it has been sought that the experience be as far as possible to use a third -party service. That said, this is what you should know about Hyperai in Europe.
- Gemini: Integration with Gemini is completely native. It works as a system assistant, it is invoked with the ignition button and, the most mold, you can read the content of all our apps if we give permission. This allows you to make summaries, translations, and others.
- Writing with AI: We can change the style of the message, make summaries, review texts, etc.
- Gallery editor with AI: We can improve the clarity of the photos with AI, make generative expansion of them (such as the adobe tool), delete objects, or create videos with AI.
- Voice recognition: Transcripts, distinction between speakers, summaries, translations and custom labels for each speaker.
- Interpret conversations with AI: Face to face interpretation, automatic language recognition, translation templates.
- CALL INTERPRETE
- Subtitles with real time
Xiaomi’s AI is quite similar to view on the rest of the phones that use the Nano Gemini model: translations, text improvements, editions in the gallery … Those that require Internet connection, such as the expansion of images, also require Xiaomi account.
The deletion of people is improvable.
Generative for edges.
After trying some of these functions, I am clear that Xiaomi’s AI is in its first steps. In this example, quite easy by the way, I wanted to eliminate the person from photography. The result is somewhat inconsistent. In the second example I wanted to fill in space in photography. If we expand, we see that elements such as the fork contain errors.
Writing functions are available in apps such as WhatsApp. Some others are native, such as interview transcription in their recorder app.
Regarding text corrections, they work quite well. It is a useful function for professional emails and work environment. Transcripts work relatively well (with some errors, of course), but the phone takes a long time to process the results. I would like this processed to be done in the background to expedite the process.
Xiaomi has reached AI a tuño late, but the integration promises. Not taking some year of experience makes it end up noticing a certain distance in front of some of its competitors, so the following updates have an important work ahead: reach the level of the best in the market.
Image | Xataka