Last week, Google announced the public preview of its Gemini 2.5 Pro artificial intelligence (AI) model. The Mountain View-based company had previously launched this flagship AI model as an experimental preview with limited usage rates just last month. With the move to public preview, users of the Gemini application programming interface (API) and those accessing the model through Google AI Studio will now benefit from increased usage limits.
Gemini 2.5 Pro Pricing Details Released
In a recent blog post, Google reported strong early adoption of the Gemini 2.5 Pro model among developers. Due to this significant interest, the company is broadening access to the large language model (LLM). The model is now available in public preview within the Gemini API on Google Studio, complete with enhanced rate limits. It’s important to note that the AI model has not yet been released on Vertex AI, but Google plans to roll it out on that platform soon.
Regarding pricing, Gemini 2.5 Pro offers two distinct tiers. The first tier accommodates regular rate limits of 200,000 tokens or fewer, while the second tier applies once the regular rate is exceeded.
Under the regular rate, Gemini 2.5 Pro will charge $1.25 (approximately Rs. 107) per million input tokens, which covers text, images, audio, and videos. The cost for output tokens is set at $10 (about Rs. 858) per million, with adjustments made for reasoning tokens.
Once the 200,000 token threshold is surpassed, developers will incur a fee of $2.50 (around Rs. 214) for each additional million input tokens and $15 (approximately Rs. 1,290) for every million output tokens. The experimental version of the model remains free but is subject to lower rate limits.
Google emphasizes that its pricing structure is “competitive.” A comparison with competitor offerings shows that Gemini 2.5 Pro appears significantly more affordable. For instance, Anthropic’s flagship model, Claude 3.7 Sonnet, charges $3 per million input tokens and $15 per million output tokens.
Furthermore, in the competitive landscape of AI, OpenAI’s pricing is even steeper. The o1 AI model sets the input cost at $15 per million tokens and $60 (roughly Rs. 5,150) for output tokens, although it does offer a discounted rate of $7.50 (approximately Rs. 640) for cached input tokens.
