Google Launches Gemini 2.5 Pro: Advanced AI Model Expands Features Across Apps and Devices
Image Credit: Dylan Carr | Splash
Google has announced the launch of Gemini 2.5 Pro, its latest artificial intelligence model, at the Google I/O 2025 conference held in Mountain View, California, on May 20–21. The new model features advanced reasoning, coding, and multimodal processing capabilities, supporting Google’s aim to strengthen its position in the global AI sector.
Gemini 2.5 Pro: Enhanced Performance and Integration
Gemini 2.5 Pro is designed to process text, images, and video, and is being integrated into several Google products, including the Gemini app, Google Search, Chrome, and supported smart devices. The model introduces a new "Deep Think" mode, which Google says enables more comprehensive analysis by evaluating multiple possible answers before responding. As of late May 2025, “Deep Think” mode is not yet publicly available to all users. It is currently in testing with a limited group of trusted users through the Gemini API and Vertex AI, with Google conducting further safety evaluations before wider release. The company has indicated that “Deep Think” will become available to consumers as part of the Google AI Ultra subscription plan following this evaluation phase.
According to Google, Gemini 2.5 Pro has achieved top results on a range of industry-standard AI benchmarks, including human preference, web development, and mathematics. Google also reports improvements in coding, science, and general knowledge benchmarks, surpassing several competing models.
A streamlined version, Gemini 2.5 Flash, offers similar capabilities with improved efficiency, using fewer computational resources to deliver faster real-time performance. Both models incorporate enhanced security measures designed to prevent emerging threats, including prompt injection attacks.
Applications Across Google Ecosystem
Gemini 2.5 Pro’s multimodal features are now available in Gemini Live, a service that allows users to analyze and interact with images via mobile devices. The model also supports functions in Chrome, such as summarizing web content and answering queries. For developers, Gemini 2.5 Pro can generate code, build web applications, and assist in content creation using simple text prompts. In the automotive sector, Google has partnered with Volvo to deploy Gemini-powered voice assistants in select car models, providing hands-free access to vehicle information. The model also powers functions in emerging technologies such as Android XR smart glasses, offering real-time navigation and translation.
Product Development and Market Competition
Google’s release of Gemini 2.5 Pro follows increased competition from companies including OpenAI, Anthropic, and x.ai. The new model builds on advances in Google’s earlier Gemini 2.0 system, and the company says it features a larger context window—capable of processing up to 1 million tokens, with plans to expand further—enabling complex analysis of lengthy documents and codebases.
Accessibility and Regional Availability
Gemini 2.5 Pro is officially available in numerous countries and territories, including Australia, through the Gemini app and Google AI Studio. However, Hong Kong is not included in Google’s official list of supported regions. Users in Hong Kong may attempt to access Gemini 2.5 Pro via VPN services and by registering a Google account in a supported country, but this method is not officially endorsed by Google and may present limitations.
Advanced features of Gemini 2.5 Pro require a paid subscription, with the Google One AI Premium plan priced at USD 19.99 per month. Google also offers an Ultra plan for enterprise users with expanded capabilities.
Some Gemini 2.5 Pro features are rolling out in stages, with certain integrations initially available only in selected regions and devices. In automotive applications, for example, deployment currently focuses on specific Volvo models.
Benefits and Limitations
Gemini 2.5 Pro supports a wide range of applications, including education, productivity, and creative content development. The model’s enhanced security features are intended to align with global data privacy standards. However, the cost of advanced subscription plans and the phased regional rollout may limit immediate accessibility for some users. In India, Google has faced trademark-related challenges over the “Gemini” name, but this has not been reported to affect other markets in Asia at this time.
TheDayAfterAI News
We are a leading AI-focused digital news platform, combining AI-generated reporting with human editorial oversight. By aggregating and synthesizing the latest developments in AI — spanning innovation, technology, ethics, policy and business — we deliver timely, accurate and thought-provoking content.