Google has opened broader access to its Gemini 2.5 Pro flagship model, introducing competitive pricing and higher usage limits in an expanded public preview. Demand for the model has surged by over 80 percent this month across Google AI Studio and the Gemini API.
- The API follows a tiered pricing structure, with input costs of $1.25 per million tokens for prompts up to 200,000 tokens and output at $10.
- On the GPQA Diamond benchmark, the model scored 84%, outperforming the typical human expert score of 70%.
- It achieved 18.8% on the "Humanity's Last Exam," significantly higher than competitors like Deepseek-R1's nine percent.
- Weekly testing on trackingAI.org showed an average IQ of 130, while users praise its cost-benefit ratio and coding capabilities.
The expanded access allows more users to utilize a model that Google describes as its most intelligent and in-demand, with additional announcements expected at the Cloud Next '25 conference.