Google was overshadowed a bit last week, releasing the new Gemini 2.5 Pro AI model while everyone was busy talking about GPT-4o image generation and its Studio Ghibli ripoffs. But Gemini Advanced users got access to the best possible Gemini AI tool Google has to offer, and paid users who did use Gemini 2.5 Pro were quick to hype the AI.
However, Google is going forward with an unexpected move: it’s making Gemini 2.5 Pro available to free Gemini users. There are some limitations, but free users will still be able to test the latest and best Gemini AI model.
Gemini said last week that Gemini 2.5 is its most intelligent AI model. “Our first 2.5 release is an experimental version of 2.5 Pro, which is state-of-the-art on a wide range of benchmarks and debuts at #1 on LMArena by a significant margin,” the company teased.
“Gemini 2.5 models are thinking models, capable of reasoning through their thoughts before responding, resulting in enhanced performance and improved accuracy.”
Google also explained that it aims to bring reasoning abilities to all its AI models with the help of Gemini 2.5. Reasoning AI chatbots will do more than “classification and prediction.” Reasoning “refers to [an AI’s] ability to analyze information, draw logical conclusions, incorporate context and nuance, and make informed decisions.”
OpenAI might have been the first to launch a reasoning model with ChatGPT o1, but rivals have been catching up. That’s why DeepSeek R1 stunned the world and tanked the AI stock market a few weeks ago. Claude also got reasoning abilities in the 3.7 Sonnet update, and now it’s time for Google to bring improved reasoning to Gemini models.
Gemini 2.5 reasoning abilities will be available in all of Google’s models going forward.
In addition to reasoning, Gemini 2.5 Pro Experimental excels at various tasks, surpassing rival models in various benchmarks. It has improved code capabilities, too, with Google demoing the creation of browser-based games from a single text prompt.
The social media reactions to Gemini 2.5 Pro have been great so far if you were not absorbed by all the Ghibli-style AI images that flooded the internet. Given all those reactions, I’ve been considering giving Gemini more attention. The good news is that I could be accessing it for free now that Google is ready to roll it out to users who do not pay the $20/month Advanced subscription.
The company announced on X that all Gemini users will be able to try out Gemini 2.5 Pro Experimental in the Google AI Studio and the Gemini app on the desktop.
Unlike Gemini Advanced users, Gemini Free users will have a few limitations. They’ll experience both rate limits and a reduced context window. It’s unclear how small the latter is, but Gemini 2.5 Pro for Advanced users has a 1 million token context window, with Google aiming to double that. Nobody else in the industry can match that. The larger the context window, the better the AI experience, regardless of the task at hand.
That said, Google didn’t specify actual limitations, so it’s up to Gemini Free users to discover them. Also, Google plans to bring Gemini 2.5 Pro to everyone on mobile, but there’s no timeline for that.
