Google Introduces Gemini 2.5 as its Most Intelligent AI Model; Gemini 2.5 Pro Experimental Debuts at #1 on LMArena
The California-based tech maker – Google has now introduced its Gemini 2.5 ‘thinking’ model. The new AI model is claimed to be the brand’s most intelligent AI model so far and its first release – the Gemini 2.5 Pro Experimental has debuted ranking first on the LMArena benchmark.
Read more about it below.
Gemini 2.5 Thinking Model Introduced
As mentioned, the Gemini 2.5 AI model has now been introduced by Google and it has been brought out as a ‘thinking’ model which will be capable of tackling increasingly complex problems. It will be able to reason before making a response, thus assuring improved accuracy and enhanced performance. An experimental version of Gemini 2.5 Pro will be the brand’s first Gemini 2.5 release.
The Gemini 2.5 Pro Experimental has topped in the LMArena benchmark, and it has showcased both strong reasoning and code capabilities – leading benchmarks of math, science, and common coding (GPQA, AIME 2025, SWE-Bench Verified). Speaking more, Gemini 2.5 Pro has scored 18.8% across different models on Humanity’s Last Exam (without tool use), and scored 63.8% on SWE-Bench Verified.
As of now, the Gemini 2.5 Pro has been made available on the Gemini app for advanced users and also on the Google AI Studio. It will also be coming soon on Vertex AI as well. About the pricing details, it will be announced in the coming weeks.
Other Updates
In related news, the brand officially launched its fifth Pixel 9 series smartphone – the Google Pixel 9a a few days back, and its sale is said to begin from next month, in April 2025 (however, it could be delayed). The WearOS 5.1 update was also rolled out to Pixel Watch, Pixel Watch 2 (LTE), and Pixel Watch 3 (LTE) recently.
