Google made several AI-centric announcements at its I/O 2026 conference, including a price reduction for its top-tier AI Ultra plan. However, the company quietly introduced significant changes to its lower-tier AI Pro plan, imposing stricter usage limits and switching to a credit-based system. This move has drawn criticism from users who feel the changes are unfair and deceptive.
What changed with AI Pro
Previously, the AI Pro plan offered a straightforward usage allowance based on token counts, with no time-based restrictions. Now, Google has adopted a new credit system that allocates tokens based on the complexity of the user's prompt, the features used, and the length of the chat. This means that a simple text query might consume fewer credits, while a complex multimodal request could rapidly deplete the daily quota. According to user reports on Reddit, a single text prompt burned through 13% of their total credit allocation, suggesting that the new system is far more restrictive than the old one.
Additionally, Google is now enforcing usage limits within five-hour windows, similar to Anthropic's approach. These windows reset after each cycle until the weekly limit is reached. Users can track their consumption via the 'Usage limit' option in Gemini’s settings. The changes also apply to other AI products under Google’s umbrella, including Antigravity and Flow, though it is unclear whether the context windows have been altered.
Comparison with the Ultra plan changes
While the AI Pro plan has been effectively nerfed, Google simultaneously announced a $50 price cut for its AI Ultra plan, reducing it from $200 to $150 per month. A new $100 per month AI Ultra plan was also introduced, positioned between the original Ultra and the Pro tiers, with fewer perks and lower limits. This tiered structure suggests Google is segmenting its offerings to capture different user budgets, but the Pro plan changes have left many feeling shortchanged.
The timing of these changes is notable. Just days before the announcements, Google had been experimenting with weekly limits on certain accounts. The new policies confirm that these limits are now permanent. Google also notified users via email, but the lack of transparency about the exact token or credit allocations has fueled distrust.
Historical context and industry trends
Google’s move reflects a broader industry shift toward more granular usage metering. Competitors like OpenAI have long employed tiered token limits, while Anthropic introduced five-hour windows earlier this year. However, the backlash against Google’s changes is particularly intense because the company marketed the Pro plan as providing 'significantly higher usage limits' compared to the free tier. Under the new system, the exact multiplier is unclear; Google only states that Pro users get four times the credits of free users, but since free tier limits are also vague, the effective value is debatable.
Historically, Google has adjusted its AI subscription plans multiple times since launching Gemini Advanced. The Pro plan originally offered a fixed number of queries per day, then migrated to token-based caps, and now to a dynamic credit model. These shifts indicate a struggle to balance server costs with user expectations. The AI arms race has intensified computing demands, and providers are seeking ways to prevent abuse while maintaining affordability.
User reactions and community response
On Reddit and other forums, users have expressed outrage. One popular post described the changes as 'a total scam,' noting that a single complex prompt consumed a disproportionate amount of credits. Others have pointed out that the five-hour window system penalizes power users who need continuous access. Some have compared Google’s approach unfavorably to OpenAI’s more transparent token usage dashboard. The lack of advanced notice has also been criticized; affected users received an email after the changes took effect, leaving them with no opportunity to adjust their workflows.
Google has defended the new system as 'a fairer approach' that charges based on actual compute rather than raw token count. However, the company has not released detailed documentation on how credits are calculated, causing uncertainty. Analysts suggest that Google is following the precedent set by cloud computing providers, where variable pricing based on resource consumption is standard. But for consumers accustomed to fixed subscriptions, the unpredictability is unwelcome.
Impact on power users and developers
The changes are likely to affect developers and researchers who rely on the AI Pro plan for building prototypes or conducting experiments. Under the old system, they could run multiple long-context sessions without hitting caps. Now, with the five-hour reset windows, they must carefully manage their usage or risk being throttled mid-project. Google also discontinued the 1,000 free AI credits for Flow, though the company claims this will not change user experience. For heavy users, the new limits might push them toward the more expensive Ultra plans or to competitors.
Another concern is the lack of guarantee for future pricing and limits. Google’s terms of service allow it to modify these parameters without prior notice, meaning users cannot rely on the current allocations for budgeting. This instability could undermine trust and prompt a migration to alternatives like OpenAI’s ChatGPT Plus or Anthropic’s Claude Pro, which, despite their own cap changes, have historically provided more consistent usage metrics.
Broader implications for AI subscription market
Google’s quiet nerf of the AI Pro plan may signal a new phase in the AI subscription market where features and limits are constantly in flux. As companies seek to monetize their AI infrastructure, they are experimenting with different pricing models. The credit system, while arguably more equitable for light users, penalizes those who need sustained high throughput. This could lead to a two-tier market: casual users might benefit from lower costs, while professionals and enterprises are forced into premium tiers with higher margins.
Google’s decision also highlights the tension between competing on price and maintaining server capacity. The price cut for the Ultra plan suggests Google is willing to sacrifice per-user revenue to attract high-value customers, while the Pro nerf squeezes mid-tier users. This strategy mirrors what telecom companies have done with data plans: offering cheap entry-level tiers but with strict throttling, and premium tiers for unlimited access.
For now, users are advised to monitor their credit consumption closely and consider whether the Pro plan still meets their needs. Alternatives exist, but each come with their own trade-offs. Google’s changes may also encourage the development of more efficient prompt engineering techniques and the use of offline or local AI models for certain tasks.
The situation remains fluid. Google has not announced any further adjustments to the Pro plan, but the backlash might prompt a revision. In the meantime, the community is sharing tips on how to maximize the credits, such as breaking long prompts into smaller chunks and using shorter context windows when possible. Whether these workarounds are sufficient remains to be seen.
Source: Android Authority News