Google reshuffles Gemini AI plans with new Ultra tier, revises usage limits

Google announced new AI subscription tiers at Google I/O 2026, including an affordable Ultra tier for developers and revised usage limits for AI Pro, shifting from fixed quotas to compute-based limits. Critics argue the changes reduce transparency, while the new Ultra plan offers 5X higher usage limits and 20TB cloud storage for heavy AI users.
Google unveiled major updates to its Gemini AI subscription plans during Google I/O 2026, introducing a new AI Ultra tier priced at $100 per month. This plan targets developers, advanced creators, and technical professionals, offering up to five times higher usage limits in the Gemini app and Google Antigravity, along with priority access to AI agent development tools and 20TB of cloud storage. The company also revised its AI Pro plan, replacing fixed monthly quotas with compute-based limits that depend on prompt complexity, chat length, and AI features used. Users reported confusion over the changes, particularly after Google removed the 1,000 monthly AI credits previously included with the subscription, forcing them to purchase additional credits separately for extended usage. Despite the criticism, Google confirmed that AI Pro subscribers will retain benefits such as access to Gemini Flash and Pro models, advanced research tools, video generation, and 5TB of cloud storage. The updated limits refresh every five hours until users reach their weekly quota, though exact token or prompt thresholds remain undisclosed. New features like Gemini Spark and Gemini Omni were also announced, alongside a revamped Daily Brief tool. The Ultra tier aims to streamline workflows for heavy AI users, including coding, media generation, and automation, while the AI Pro adjustments reflect Google’s shift toward dynamic resource allocation. User feedback on social media highlighted concerns over predictability and transparency in the new system, contrasting with the previous fixed-quota model. Google has not provided detailed breakdowns of the compute-based limits, leaving some subscribers uncertain about how usage will be tracked moving forward.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.