Following changes to the Gemini app in May, Google is updating Gemini Notebook with a new system of compute-specific usage limits that will take effect in September. This approach replaces the previous daily feature-generation caps that varied by Google AI tier and introduces a more nuanced way to measure and manage AI usage across Notebook features.
- Related: Google announces ‘Expert Intelligence’ to let you add Play Books in Gemini Notebook
Instead of a fixed count of feature generations per day, the new system evaluates each request based on the compute resources it is expected to consume. Limits refresh every five hours rather than once per day, giving users more frequent replenishment of their available compute budget and more flexibility in how and when they use Notebook capabilities.
Google notes that a user’s overall usage limit will account for multiple factors, including prompt complexity, chat length, the number of sources referenced, and the particular features in use. Notebook will surface clear status information beneath the chat to help you monitor consumption, with messages such as:
- “You’re almost at your AI usage limit. Limit resets at 3:00 PM.”
- “Limit reached. All features are available after 3:00 PM.”
In addition to these status messages, Studio generation controls will show an indicator at the bottom of the generation panel that represents the expected AI usage cost. The interface will visually convey that “the more the bar is filled in, the higher the expected cost,” helping users make informed choices before initiating resource-intensive outputs.
To reduce friction when a requested output would exceed your current compute allocation, Gemini Notebook will propose alternate, lower-cost outputs. There will also be a “Generate later” option allowing you to defer the creation of heavy outputs—such as detailed Video Overviews or comprehensive Slide Decks—until your compute limit resets. When you choose to defer, Google will automatically generate those outputs once your allowance renews. Initially, the “Generate later” capability will be available on the web, with wider rollout expected over time.
- Free/Without a plan: Standard compute limits
- Google AI Plus: Approximately 2× the standard limits
- Google AI Pro: Approximately 4× the standard limits
- Google AI Ultra: Between 5× and 20× the AI Pro limits depending on the specific Ultra subscription
These tiered multipliers mean that subscribers receive proportionally larger compute budgets aligned with their plan level, making it easier to run more complex or frequent Notebook tasks. The exact behavior for any given request will still depend on the elements listed above—prompt length and complexity, the number of referenced sources, and which Notebook features are used—so users on higher tiers will enjoy more headroom for demanding workflows.
Google has said the compute-based usage limits will begin rolling out to consumer accounts on web and mobile starting September 2. The revised cadence and the more granular cost indicators aim to provide users with clearer feedback about their AI usage, encourage more efficient requests, and reduce unexpected interruptions when working with Notebook features.
Overall, this shift from simple daily quotas to compute-aware limits is intended to offer more predictable and equitable access to Notebook capabilities across different user tiers, while giving everyone better tools to manage their AI consumption in real time.