Google has significantly restructured its artificial intelligence offerings, unveiling a rebranded AI subscription service and a new top-tier plan. The company’s “AI Premium” is now “Google AI Pro,” maintaining its $19.99 monthly US price, while the more powerful “Google AI Ultra” debuts at $249.99 per month. This overhaul redefines user access to Google’s advanced AI, including its Gemini models, and signals a clearer monetization strategy.
The revamped Google AI Pro plan continues to provide 2TB of cloud storage, Gemini 2.5 Pro model access, Deep Research, and Veo 2 video generation. In a significant step up, Google AI Ultra encompasses all Pro benefits but adds 30TB of storage, a YouTube Premium individual subscription, higher usage limits, and exclusive access to forthcoming features like the 2.5 Pro Deep Think mode and Veo 3 video generation, according to Google One’s official plans page. Ultra subscribers also get early access to Project Mariner, an AI agent prototype. The previous “Gemini Advanced” branding is now discontinued.
This restructuring aims to provide distinct value for various users, from students with special access programs to professionals needing maximum AI capabilities, as outlined on the Gemini subscriptions page. The Google AI Ultra plan, initially available in the US with an introductory price of $124.99 for three months, will roll out to other countries soon. Shimrit Ben-Yair, Vice President of Google Photos and Google One, stated that the Google AI Ultra plan is “built for you — think of it as your VIP pass to Google Al” for those who “demand the absolute best of Google AI.”
Diving Into The AI Tiers: Pro And Ultra
Google AI Pro, at $19.99 per month in the US, offers enhanced access to Google’s premium AI. This includes the Gemini 2.5 Pro model, tools like Deep Research, and Veo 2 for AI video creation. Pro subscribers also benefit from Gemini integrations in Workspace apps, NotebookLM Plus features, and the Whisk tool.
New to the Pro tier are early access to Gemini in the desktop Chrome browser and the Flow AI filmmaking tool (both initially US-only), the latter using Veo 2 with a 100-generation monthly limit. Eligible university students in countries like Japan and the UK can access Google AI Pro free for an academic year through the student discount program.
The Google AI Ultra plan, priced at $249.99 monthly, represents Google’s most comprehensive consumer AI package. It includes a substantial 30TB of storage and a YouTube Premium plan. Ultra users are promised the “highest limits” for AI tools, including the upcoming Veo 3 model in Flow, which will also feature native audio generation.
Access to an experimental “Agent Mode” and the advanced “2.5 Pro Deep Think” mode for the Gemini app are key differentiators, though Google notes some of these advanced features are “coming soon.” This pricing places AI Ultra $50 above competitor plans like ChatGPT Pro and Claude Max. The free tier of Google AI continues to offer significant features, and for those interested in the expanded capabilities, Google is offering a one-month free trial for the AI Pro plan.
Advanced Capabilities And The Road Ahead
Central to both paid tiers is access to Google’s powerful Gemini AI models. Gemini 2.5 Pro, known for its 1 million token context window (allowing the AI to process and ‘remember’ a significantly larger amount of information in a single interaction), underpins the Pro plan.
Advanced multimodal reasoning in Gemini 2.5 Pro enables understanding of visual layouts in PDFs. Google’s own developer documentation refers to this as “native vision,” though official Google Cloud documentation also cautions that its models “aren’t precise at locating text or objects in PDFs,” adding that they might only provide approximate counts.
Video generation is a key focus in Google’s strategy. Veo 2 is available in the Pro tier, enabling users to create video clips from prompts via tools like “Whisk Animate”. The Ultra tier will introduce Veo 3, promising higher quality video and audio generation, including dialogue and background noise.
Google also confirmed that its SynthID technology is used by default for Veo, Imagen, and Lyria models to embed an invisible watermark for identifying AI-generated media, according to a Google Cloud Blog post. Deep Research capabilities are available in Pro and enhanced in Ultra.
NotebookLM also receives tiered improvements, with Pro users getting expanded limits and Ultra users anticipating the “best model capabilities” in the future.
Project Mariner, an AI agent available to Ultra users, leverages advanced vision-language models to interpret visual elements and text, capable of handling up to 10 simultaneous tasks and featuring a “Teach and Repeat” capability. Google stated in its blog that “We think of agents as systems that combine the intelligence of advanced AI models with access to tools, so they can take actions on your behalf and under your control.”
These agentic capabilities are also being brought to developers via the Gemini API and will be integrated into Google products “throughout the year,” according to Google
The refined subscription model arrives as Google makes substantial investments in AI, including, as revealed in antitrust trial testimony, paying Samsung “enormous sums of money” for Gemini preinstallation. This strategy, part of a broader AI integration across Google’s ecosystem, aims to provide users with enhanced productivity and creativity. Google CEO Sundar Pichai said at Google I/O 2025, “The opportunity with AI is truly as big as it gets.”
