New more affordable AI plan revealed for Google Gemini

It has been revealed that a new subscription layer has been prepared for Google Gemini services. Codes detected in the latest version of the company’s Gemini application published for macOS point to a new option called “AI Ultra Lite” that will be positioned between the existing $ 20 Pro plan and the $ 250 AI Ultra plan. In addition, it seems that Google is also working on a special usage panel where users can view their token usage amounts and remaining limits. Although there is no official announcement yet, the emerging details show that Google is making a new preparation, especially for the intensive use of artificial intelligence.

While there has been a significant increase in the use of productive artificial intelligence tools recently, code generation and agent-based systems, in particular, have led service providers to impose new limitations due to high token consumption. While users on Anthropic’s Claude service stated that usage limits had decreased, GitHub Copilot also changed the pricing structure for some models. This situation caused heavy users to turn to more expensive subscription packages. The new plan prepared by Google is expected to meet a similar need.

The Gemini ecosystem continues to expand recently with various tools such as AI Studio, Gemini CLI, Stitch and Antigravity IDE. Despite this, Google’s artificial intelligence-supported code generation tools have not found as strong a response in the developer community as their competitors. However, it is stated that the company has accelerated its work in this field before the upcoming Google I/O 2026 event. In a recent news published by The Information, it was reported that Google co-founder Sergey Brin led a special team to improve Gemini’s code generation capabilities.

Google AI Ultra Lite plan may target mid-range users

According to data detected in the macOS version of the Gemini application, the new subscription plan is currently being developed under the code name “Neon”. The naming seen in the application is “Google AI Ultra Lite”. However, it is stated that this branding may not be the final name. The company has not yet shared any information about pricing. Despite this, industry expectations indicate that the new plan can be positioned at a level between 50 and 150 dollars per month. Considering that OpenAI and Anthropic offer $100 subscription options, it seems likely that Google will opt for similar pricing.

For users who currently use Google AI Pro subscription and need higher processing capacity, the options are quite limited. Users either have to switch directly to the Gemini API and pay consumption-based, or upgrade to the $250 monthly AI Ultra package. This price difference is quite high, especially for individual developers and small teams. The new Ultra Lite plan may be intended to fill this gap.

In addition, the AI ​​Ultra subscription already offers access to experimental tools such as Project Genie, higher usage capacity in the Flow video and music production tool, larger notebooks on the NotebookLM side, and additional premium features in various Google services. It is not yet known what additional features the AI ​​Ultra Lite plan will come with. The information that has emerged so far shows that the most significant difference will be increased usage limits.

Another innovation that Google is working on is a special panel that shows token usage status. In this way, users will be able to directly see their current usage amounts, remaining limits and reset times. A similar system already exists on Anthropic’s Claude platform. Claude users can track their token budget in the current session, weekly limits and the next reset time in detail.

The new usage screen on the Gemini side is expected to offer a similar structure. The current system used on the company’s Antigravity platform has five-hour and weekly usage limits. In addition, there is an additional credit system for users to exceed the current subscription limit. Thanks to visual usage indicators, it is aimed that users can evaluate their current capacities more clearly before starting code generation sessions that consume intensive tokens.