Google appears to be preparing a new mid-tier subscription called “AI Ultra Lite” to sit between its $20-per-month Pro plan and the $250-per-month Ultra tier, while also adding a dedicated dashboard that shows subscribers their remaining token budget. These changes signal a response to rising demand and growing concern about token limits as more users rely on AI agents and coding tools.
Over the past several months, heavy use of AI-powered coding assistants and multi-turn agents has produced what many users describe as a token crunch. Some users of competing services have reported tightened per-session limits, and major developer tools have adjusted pricing to account for heavy model usage. In that context, many subscribers have considered or already moved to higher-priced plans to preserve uninterrupted access to high-capacity models.
Google’s Gemini ecosystem provides a range of tools for developers and creators, including a command-line interface, IDE integrations, and AI studio features. Despite this, Gemini’s coding capabilities have not yet dominated developer workflows in the way some competing systems have. With Google’s annual developer conference approaching, the company appears to be positioning Gemini to better serve power users and developers—especially those who rely on token-intensive workflows like agent-driven coding sessions.
Reports suggest Google has formed a focused team to improve Gemini’s code generation quality and overall developer experience. If those improvements significantly increase demand, more subscribers could intensify their use of Gemini for coding, which would in turn push many beyond the usage limits of current mid-tier plans. At present, options to support substantially heavier usage are limited: customers can either pay directly for API usage or move to the top-tier Ultra plan, a jump that represents a steep price increase compared with the Pro tier.
Recent app files for Gemini on macOS reveal an in-between plan with the internal codename “Neon,” surfaced under the working name “Google AI Ultra Lite.” While the final branding and pricing have not been announced, an intermediate tier would likely fall somewhere between existing Pro and Ultra prices. Industry peers offer $100 plans, so a price in that range would be consistent with competitive positioning and would bridge the gap for users who need more headroom than Pro but do not require the full Ultra feature set.
Details about the specific feature differences between Pro and the proposed Ultra Lite tier are sparse. The full Ultra plan currently includes exclusive demos, expanded access to the Flow multimedia generator, larger NotebookLM sessions, and premium integrations across a number of Google apps. Ultra Lite would most likely focus on increased usage allowances and potentially modest feature upgrades tailored to developers who need more tokens for sustained coding and multi-step agent workflows.
In addition to a potential new subscription tier, Gemini’s app references a dedicated usage page intended to show token consumption and limits. This usage dashboard appears designed to help users understand their available session tokens, weekly budgets, and any overage credits that may apply. A visible usage indicator gives subscribers a practical way to decide whether they can start a token-heavy coding session immediately or should defer until their allowance resets.
Usage limits
gemini.google.com/usage
The planned usage page mirrors the transparency some competitors provide, making session and weekly consumption clear and predictable. Internally referenced labels suggest separate five-hour and weekly quotas plus mechanisms for overage credits, which would help smooth workflows for people running longer agent sessions or completing large code generation tasks.
GXU_FIVE_HOURLY
GXU_WEEKLY
OVERAGE_CREDITS
Combined, an upper-mid-level subscription and an explicit usage dashboard indicate Google is preparing for increased and more predictable demand for Gemini and its AI offerings. Whether Google can convert that preparation into broader adoption will depend on the quality of upcoming code generation improvements, the specific features included in any new tier, and how pricing aligns with the needs of developers and heavy users.
For now, subscribers and prospective customers should watch for official announcements that confirm pricing, exact feature differences, and the launch of the usage dashboard. Those components will determine whether AI Ultra Lite becomes a practical bridge between casual use and the resource-heavy workflows that currently require a large jump to Ultra or direct API billing.