Everything Google announced at I/O 2025: AI is advancing on all fronts

🧠 Gemini 2.5: Google's new intelligence

Gemini's updated lineup of AI models has two major players:

Gemini 2.5 Pro with a Deep Think* mode that allows models to hypothesize and analyze complex problems (e.g. code and math).

* Gemini 2.5 Flash is a simplified version with almost the same quality, but 20-30% more economical in terms of tokens. Ideal for developers and budget solutions.

Google claims that Pro now leads in all LMArena categories, overtaking even GPT-4.

🗣 Project Starline became Google Beam

3D video calling technology is now beyond experimental.

Google Beam is a new business device based on a 6-camera system and real-time AI image reconstruction.

📡 Running at 60 FPS with millimeter head tracking.

📦 First commercial release is from HP, customers: Salesforce, Deloitte and Duolingo.

💸 AI Ultra - $250/month subscription.

Google has introduced premium rate AI Ultra:

* ♪ $250 a month (first three months $125) ♪

Unlimited access to new AI features (including Deep Research)

30 TB of storage in Photos, Drive and Gmail

* YouTube Premium and access to experiments like Project Mariner.

🔍 New Search with AI Mode

AI Mode is a separate search mode that turns Google Search into a chat with AI.

Features:

*Deep Queries and Sequential Refinements.

* Deep Search: up to hundreds of background queries per answer.

* AI-assisted shopping with photo-based fitting.

Available starting this week in the US.

🕶 Android XR and Project Aura - new augmented reality glasses

Google has unveiled the Android XR platform and prototype Project Aura glasses:

* In conjunction with Xreal, support for Gemini.

* Support for navigation, translations, 360° video.

Design partners: Samsung, Warby Parker and Gentle Monster

📩 Gmail will learn to write for you

Coming soon to Gmail:

Personalized smart replies that adjust to your communication style

Gemini will be able to use documents from Drive and mail to compose replies

Launch: July via Google Labs

🌍 Google Meet now translates in real time

• Real-time speech translation with save intonation
• English and Spanish support at launch
• First users are AI Pro and Ultra subscribers

🤖 Project Mariner - AI with memory and automation

• Supports up to 10 simultaneous tasks
• Teach & Repeat: you can teach Mariner to perform actions by example
• API already available for partners (UiPath, Automation Anywhere)

🧭 Project Astra - universal AI assistant

Google has revealed Astra, an advanced assistant with visual capabilities:

• Uses smartphone camera to "see" the world
• Can look up parts, call stores, suggest repairs
• Already in Gemini Live on Android, iOS in the process of release

🎬 Veo 3, Imagen 4 and Flow - the art of content generation.

• Veo 3 - video generator with voiceovers, sounds and music
• Imagen 4 - improved generation of photorealistic images (up to 2K)
• Flow - mobile app for creating 8-second AI clips and assembling scenes

✅ Output

Google has shown that AI is not the future, but the present. From texting to video calling, from video generation to new forms of search, everything is permeated with Gemini technology.

And while the Ultra subscription price tag is still biting, the trend is clear: AI will become more and more deeply integrated into our everyday lives.

🚀 Join NakMo!

NakMo is more than just news. It's a platform where everyone can be an author:

• Share your opinions on equal footing with journalists and tech-enthusiasts
• Publish articles and they can be featured on the main page.
• Contact us via the form on the website or directly

Be on topic. Be with NakMo.