Google Offers Higher Rate Limits in AI Studio and Moves Gemini 2.5 Pro to Public Preview

Last week, Google made the Gemini 2.5 Pro AI model available for public evaluation. Last month, the tech giant based in Mountain View unveiled its most recent top tier AI model as an experimental preview with restricted rate restrictions.

However, users who access the model using Google AI Studio or the Gemini application programming interface (API) will now be subject to stricter rate constraints. Additionally, the business disclosed the higher rate limit price, claiming that it keeps them competitive.

Google stated in a blog post that it saw developers embrace Gemini 2.5 Pro early on, and as a result of this strong demand, it is now extending developers’ access to the large language model (LLM). The tech behemoth said that a public preview of the foundation model with increased rate restrictions is now accessible through Google Studio’s Gemini API. Notably, Vertex AI has not yet made the AI model available, but the business stated that it will do so soon.

Regarding the cost, there are now two price categories for the Gemini 2.5 Pro. Regular rate limitations of two million tokens or fewer apply to the first. Once a developer has used up the standard rate limit and exceeds it, the second pricing tier takes effect. Gemini 2.5 Pro will cost $1.25 (about Rs. 107) per million input tokens and $10 (approximately Rs. 858) per million output tokens under the usual rate limit. The output price is modified for reasoning tokens, while the input tokens consist of text, photos, audio, and videos.

The tech giant claims that the model’s pricing is “competitive,” and the Gemini 2.5 Pro model does seem to be substantially less expensive when compared to its rivals. For example, Claude 3.7 Sonnet, Anthropic’s flagship model, costs $3 for every million input tokens and $15 for every million output tokens.

OpenAI, Google’s main competitor in the AI market, charges considerably more for its reasoning models. $15 per million input tokens and $60 (about Rs. 5,150) per million output tokens are the input and output prices of the o1 AI model. Nonetheless, the business charges a reduced $7.5 (about Rs. 640) per million and accounts for cached input tokens.