Google has finally lifted the veil on exactly how much its Gemini AI can be used at different subscription levels. Until now, the company only described access in vague terms like “limited” or “expanded,” leaving many users uncertain about what they were actually getting. But in a recent update to its Help Center page titled “Gemini Apps limits & upgrades for Google AI subscribers,” Google has provided hard numbers for prompts, images, and other features across tiers.

For free users, Gemini 2.5 Pro access comes with significant restrictions. They are limited to just five prompts per day on the model. On the creative side, free accounts can generate or edit up to 100 images monthly and are allowed five Deep Research reports per month. Interestingly, some tools remain universally available regardless of plan—every user can enjoy 20 Audio Overviews per day at no extra cost.

Those who opt for the paid Google AI Pro plan at $19.99 per month get far more freedom. Pro subscribers can send 100 prompts per day to Gemini 2.5 Pro, create 1,000 images monthly, and run 20 Deep Research reports daily. On top of that, they can generate three Veo 3 Fast videos—a perk not available to free users.

At the top of the ladder sits the Google AI Ultra plan, carrying a hefty price tag of $249.99 per month. This package is designed for power users, offering 500 prompts daily, the same 1,000-image monthly limit, and five Veo 3 video generations per day. Ultra subscribers also gain access to 200 Deep Research reports monthly and 10 Deep Think prompts with an expanded 192,000-token context window, making it a suitable choice for large-scale or complex tasks that demand extended context and reasoning power.

Google has clarified that integrated tools such as Canvas, Gems, and Storybook will continue to function within these boundaries, meaning the limits depend entirely on the subscription tier selected.

While some of these numbers had already been speculated by the AI community, this marks the first time Google has spelled them out in official documentation. The move provides clarity, helps users plan their AI usage, and underscores the stark contrast between free and premium offerings.

Adding intrigue to the update are fresh signs that Google may be preparing to roll out its next-generation AI model. According to reports from Testingcatalog, references to a “Gemini Beta 3.0 Pro” have recently been discovered in the Gemini CLI tool codebase, which is partly maintained by Google employees. This discovery has fueled speculation that DeepMind is already testing a successor to Gemini 2.5 Pro.

Of course, the presence of Gemini 3.0 in code does not guarantee an imminent launch. But the timing—coinciding with Google’s effort to formalize user limits—has not gone unnoticed. Industry watchers believe Google could be laying the foundation for its next big AI reveal. For now, users finally have transparency about their current Gemini privileges, while also wondering if a major update is just around the corner.