The US tech giant Google has announced a preview of its enhanced Gemini 2.5 Pro, labeling it the company’s “most intelligent” model. This upgraded version is an improvement over the Gemini 2.5 Pro introduced at the I/O developer conference in May. It delivers superior performance across key benchmarks, according to Tulsee Doshi, Google’s senior director of product management.
Google’s new and improved Gemini 2.5 Pro model did well in tests. It jumped 24 points to a score of 1470 on LMArena, a test for all kinds of tasks, making it the best. It also climbed 35 points to 1443 on WebDevArena, which checks web development skills, taking the top spot.
As per Google’s blog post, its latest and most advanced AI model, Gemini 2.5 Pro, has led the Aider Polyglot test too, proving it’s excellent at coding in many languages. Plus, it’s among the top performers on tough tests like GPQA and Humanity’s Last Exam, which look at reasoning, math, science, and general knowledge. Developers can start using this upgraded model through Google AI Studio and Vertex AI, and it’s available in the Gemini app.
It’ll be fully ready for everyone in a few weeks. The firm has made these changes based on feedback from developers. The release shows that the firm is working hard to make its AI better. This release is part of Google’s effort to keep up in the fast-changing AI world, where OpenAI’s ChatGPT is super popular. Google is focusing on being outstanding at coding and scientific thinking.
Last month, they gave a sneak peek of Gemini 2.5 Pro, which was much better at coding and came out early because developers were so excited.
Gemini is Google’s main AI family, built to handle text, images, code, and video. When Gemini 2.5 was first shown in March, it was called a “thinking” The model is effective because it can reason, rather than merely guessing or sorting data. This upgrade is another move by Google to be a leader in AI.
Also Read: OpenAI Seeks Indian Partnership Under ‘OpenAI for Countries’ Plan
