14 Januari 2026
Google I/O 2025 has set a revolution in motion. This year, the company has chosen a bold approach course: no more small steps, but a complete revision of generative AI that is reshaping the landscape for search engines, video and digital creativity.
At the heart of these changes is Gemini, Google’s next-generation model that powers everything from search results to video synthesis and high-quality image creation. At a time when speed and natural AI generation are the benchmark, Gemini positions itself as a game-changer.
The real showstopper? Veo 3, Google’s first-ever AI video creator that not only creates visuals but also generates complete soundtracks—from ambient sounds to conversations—perfectly synchronized with the visuals. Enter text and visual instructions, and voilà: a fully produced 4K video appears! This marks the first large-scale video engine capable of generating audio and visuals simultaneously, a trend that began with the not-yet-released Showrunner Alpha. Veo 3 offers significantly more versatility, capable of generating a variety of styles beyond simple 2D animations. “We’re entering a new era of creation with combined audio and video generation,” said Josh Woodward, VP of Google Labs, at the launch. With this, Google is challenging the established video leaders and positioning Veo as an all-in-one solution.
Alongside Veo 3, Google is also introducing Imagen 4, the latest version of its image generator. This version brings improved photorealistic renderings, 2K resolution, and a crucial feature: text generation that’s actually usable for signage, products, and digital mockups. For those who’ve previously struggled with AI’s jumbled text, Imagen 4 is a significant step forward.
These tools don’t stand alone, however. Flow AI, a new subscription feature for professional users, combines Veo, Imagen, and Gemini’s language capabilities into an integrated filmmaking and scene-editing environment. That convenience comes at a price of $125 per month, during a promotional period until the full $250 is charged.
Generative AI isn’t just for content creators. Gemini 2.5 now forms the backbone of Google’s new search engine, evolving from a link aggregator to a dynamic, conversational interface that tackles complex questions and delivers aggregated, multi-layered answers. AI feeds—where Google Gemini attempts to provide comprehensive answers without requiring users to click to other sites—now sit at the top of search pages. And that’s not all: with over 1,5 billion monthly users, the growth is spectacular.
Then there’s “Gemini Diffusion,” built on technology pioneered a few months ago by Inception Labs. While the AI community has previously focused on autoregressive technology for text generation and diffusion technology for images, Google has now taken the first big step by applying diffusion generation to text. This means that the model starts with random information and refines it with each iteration, at an incredibly fast speed.
Our experiences with several of Google’s new AI features were mixed, depending on the level. Deep Research proves to be especially powerful, even better than comparable alternatives. It can analyze hundreds of sources and deliver reliable data with minimal errors, even generating visually appealing infographics.
Veo 3 is currently only available to Gemini Ultra users, but some third parties are already offering access via APIs. Flow, the intuitive video editor powered by Veo’s models, allows users to easily edit AI scenes with text prompts.
As for audio feeds, the new models can now effortlessly produce over twenty minutes of full debates without users having to switch to other platforms. This makes Gemini a more complete interface, reducing the fragmentation that previously encouraged switching between different sites for different services.
Google hasn’t kept its monetization strategy a secret. The “Ultra” plan costs $250 per month and prioritizes the most powerful models, Flow AI tools, and 30 terabytes of storage. This is clearly aimed at filmmakers, serious content creators, and enterprises. For a broader audience, there’s the $20 “AI Pro” tier, which offers access to the earlier Veo 2 model and imaging and productivity features. Basic generative tools remain free, but with restrictions such as a token limit and a maximum of ten studies per month.
This layered approach reflects a broader trend in the AI market: Draw the masses to the platform with free tools, then lock in professionals with features too valuable to ignore. Google’s bet is that the real wins will lie in high-end creative work and automated enterprise workflows—not just casual prompts and meme generation.
“And now, get started with your own creative explosion?”
What is the most important innovation of Google I/O 2025?
The launch of Gemini and Veo 3, which together set a new standard for AI-generated video and visuals.
How does Gemini Diffusion work?
Gemini Diffusion starts with random information that is refined step by step, allowing the model to generate thousands of tokens per second without compromising accuracy.
How much does Google's Ultra plan cost?
The Ultra plan costs $250 per month and offers access to the most powerful AI tools and 30 terabytes of storage.
Bitcoin (BTC)
Ethereum (ETH)
XRP (XRP)
