Google's Gemini 2.5 Pro Deep Think Rolls Out to Ultra Users with Bronze-Level Performance and Strict Prompt Limits

Google has initiated the rollout of its Gemini 2.5 Pro model, featuring the advanced "Deep Think" enhanced reasoning mode, to Google AI Ultra subscribers. This powerful AI tool, designed for highly complex tasks, is currently accessible with a significant limitation: users are restricted to five prompts every 12 hours. The rollout has prompted discussions among the AI community regarding the cautious approach, especially for a service costing $250 per month.

The "Deep Think" mode is an enhanced reasoning capability within Gemini 2.5 Pro, utilizing parallel thinking techniques to tackle intricate problems. While an advanced version of Gemini Deep Think achieved a gold-medal level performance on problems from the International Mathematical Olympiad (IMO), the version released to Google AI Ultra subscribers currently demonstrates "Bronze-level performance" on the 2025 IMO benchmark. This distinction highlights different tiers of the technology's availability.

The restrictive prompt limit has drawn public scrutiny. Social media user Haider. voiced this sentiment, stating, > "still don't understand why Google hasn't released Gemini 2.5 DeepThink, even for pro users... for Ultra users who pay $250/m, DeepThink has a limit of 5 prompts every 12 hours." This perspective underscores a perceived disconnect between Google's vast resources—including compute power, distribution networks, infrastructure, and top researchers—and the limited access provided to paying users.

Google's official model card for Gemini 2.5 Deep Think indicates that the company is taking a precautionary approach to its release. Ongoing evaluations under the "Frontier Safety Framework" assess potential risks, particularly in areas like chemical, biological, radiological, and nuclear (CBRN) information and cybersecurity. These safety considerations and the implementation of mitigations may contribute to the measured rollout and current usage limitations.

The development of Gemini 2.5 Pro Deep Think represents a significant step in AI's problem-solving capabilities, excelling in areas like advanced coding and mathematical reasoning. As Google continues to refine the model and address safety protocols, the future may see expanded access and potentially higher performance tiers for a broader user base, reflecting the continuous evolution of advanced AI systems.