Gemini App Now Uses Compute-Based Limits Instead of Message Caps, Users Say Usage Drops Sharply

Google is revising the amount of usage allowed for the Gemini app. Following the implementation of the weekly usage cap and usage dashboard last week, Google has now opted for the compute-based cap. As such, it is discontinuing the previous message-based cap method; this implies that your usage will be measured based on token usage, which may significantly lower the usage capacity. Users began noticing their limits being reached much faster within hours after the changes were implemented.

Google informed users about these developments via email communication and not through the blog or social media channels. Gadgets 360 also received this email, which mentioned that the Gemini app will now use a "compute-based usage limit that takes into account the complexity of your prompt, the features you use, and the length of your chats."

The email also officially mentioned the addition of a weekly limit. However, the email did not specify any standard metrics for compute consumption for per-response or per-media generation. On a support page, Google only broadly mentions the different usage tiers: Standard Limit (free tier), 2x Standard Limit (AI Plus), 4x Standard Limit (AI Pro), 5x Standard Limit (AI Ultra $100), and 20x Standard Limit (AI Ultra $200).

Since the company isn't providing any official measurements, the publication used a chatbot to try to determine how much usage actually results from using "premium models and features." In testing, they found that generating a single Nano Banana Pro image consumed one percent of a 5-hour usage limit, while conducting deep research using Gemini 3.1 Pro (Extended) consumed five percent of a 5-hour usage limit. Generating a single video using Veo 3 consumed 26 percent of the five-hour limit. This usage consumption was observed on a Gemini account with an AI Pro subscription.