Gemini 2.5 Pro is Google’s most expensive AI model yet

Friday, Google has been released API price For Gemini 2.5 ProAn AI argument model with art-redempted performance in several criteria of coding, logic and mathematics measurement.

For the request of up to 200,000 tokens, the Jemini 2.5 Pro is a million input tokens for the Pro -$ 1.25 (about 750,000 words, the whole “lord of the ring” and millions of output token per million token. For prompts for more than 200,000 tokens (which do not support most of Google), the Gemny 2.5 Pro spends $ 2.50 per million input tokens and $ 15 per million output tokens.

This price makes Gemi 2.5 PRO more expensive for developers than any other AI model paid by Google with Google, including Gemi 2.0 flash ($ 0.10/million input token). It makes the Geni 2.5 prone more than the AI ​​model, such as Openai and 3-minute ($ 1.10/million input token, $ 4.40/million output token) and DEPSK R1 ($ 0.55/million input token, $ 2.19/million output token).

Just to say, Gemi 2.5 Pro, which is available for free with a strict rate limit, is cheaper compared to the ethnic clad 3.7 sonnet ($ 15/million output token) and OpenAE’s GPT -4.5 ($ 75/million input tokens), including $ 150/million. The initial response to the technology industry Has originally been positiveAppreciates the developers that they seem to be intelligent rates.

However, extensively, flagship models seem to have some ward -oriented pressure to determine the price. The cost of publishing recent top-line from Google, OpenAI and ethnographic national labs is not down, not below. For example, the recent launched and 1-Pro offer in the opening, which is the company’s most expensive API offer is still on $ 150/million input token and $ 600/million output tokens.

It may be that high demand and computing costs are running the trend. According to Google’s CEO Sunday Pichai, Gemi 2.5 Pro is the company’s largest in-demand AI model among developers, which leads Increases 80% in use on Google’s AI Studio platform and Gemi API this monthThe