Details on pricing and features of Google's AI packages.

At the I/O 2026 conference, Google implemented a comprehensive overhaul of its AI service offerings, launching a more affordable Ultra version for developers and integrating a range of automation features into the Workspace ecosystem.

After two years of fierce competition in AI, Google has officially redefined its product portfolio by deeply decentralizing its Gemini service packages.

The biggest highlight is the launch of the AI ​​Ultra package, priced at VND 2,250,000 per month, positioned between the expert and high-end user segments, directly competing with rivals in the programming support and digital content creation sector.

This "overhaul" introduces users to an ecosystem of four options, with pricing and benefits clearly defined based on their computing performance needs.

Details of the Google One packages currently offered. (Screenshot)

In addition to price changes, Google also introduced pivotal technologies:

1. Gemini Spark: The "sleepless" assistant

The biggest highlight is the introduction of Gemini Spark. Unlike traditional chatbots, Spark runs on separate Google Cloud virtual machines (VMs), allowing it to continue performing complex tasks even when you've closed your laptop or disconnected.

Spark can automatically browse the web, manage emails, scan credit card bills for hidden charges, or track children's school schedules to send daily summaries to parents.

Users can assign tasks to Spark via a separate Gmail address, just like assigning tasks to a colleague. This feature will launch in beta in the US for Ultra users next week.

2. Project Genie: Conjuring virtual worlds from text

Originally a test project, Project Genie has now officially expanded to include the highest-tier Ultra package. This tool allows you to create mini-games or 3D virtual environments simply by describing them with text or images.

The breakthrough is the integration of Google Street View, which allows you to bring real-life scenes into your virtual world to increase realism.

3. Gemini Omni and 3.5 Flash: Speed ​​and Multimodality

Google also updated the two core models for all paid users (Plus, Pro, Ultra).

Gemini Omni is a multimodal model that allows the creation of high-quality videos from a combination of text, images, and audio. Omni has the ability to understand physical laws such as gravity and kinetic energy, helping to create more realistic videos.

Gemini 3.5 Flash replaces version 3.1; this model is optimized for programming and agent execution tasks, with speeds up to four times faster than competitors in the same segment.

4. AI Inbox and YouTube Premium Lite

US users of AI Plus and AI Pro plans can now use AI Inbox in Gmail. This assistant not only summarizes emails but also automatically composes replies, finds relevant attachments in Docs/Sheets, and suggests to-do lists.

Additionally, Pro plan users in certain regions will receive YouTube Premium Lite as a bonus, which removes ads in content categories such as gaming, fashion , beauty, and news.

Change the method for calculating usage limits.

An important note for users is that Google has switched from counting the number of prompts to a computationally intensive model. This means that a complex video creation request will consume more resources than a simple text query.

If the limit is reached, users will be downgraded from larger AI models to smaller ones, unless they purchase additional "AI credits" individually.

According to experts, Google's decision to lower the price of its top-tier Ultra package and launch the new AI Ultra version is a move to "retain" professional users in the face of pressure from OpenAI and Anthropic.

For individual users, the AI ​​Plus package (132,000 VND/month) is currently the optimal choice for maintaining the presence of artificial intelligence in daily work at the most reasonable cost.

From now on, you can tell if an image was created by AI right on Google . Users will soon be able to directly ask Google Lens or Gemini to determine if a photo is real or AI-generated, in the context of increasingly sophisticated fake media.

Source: https://vietnamnet.vn/chi-tiet-gia-va-tinh-nang-cac-goi-google-ai-dang-cung-cap-2519376.html