Top AI news of 2025: Biggest announcements that defined the year

2025 has been big on Artificial Intelligence (AI) with fast-paced launches, new generation AI models, and more that pushed the technology closer to everyday life. Among several tech giants, Google took a major leap this year with the Gemini 3 model, showcasing significant upgrades in reasoning, multimodal understanding, and real-time productivity across the Gemini app and its ecosystem.

On the other hand, OpenAI also took a competitive stand with GPT-5 and GPT-5.2 that offered faster responses, improved contextual understanding, and other AI-powered tools as well. Beyond these major updates, the year also saw greater advances in AI-powered, creative tools, on-device intelligence, and more. 

Top AI News of 2025

Google Gemini 3.0: This year, Google unveiled its most powerful and intelligent AI model yet, the Gemini 3. The new AI models come with deeper reasoning, improved multimodal understanding, and deeper integration across Google Search, Workspace, and Android. Why is Gemini 3 hype? Well, the AI models bring advanced agentic capabilities that can help plan multi-step tasks and interact across apps more seamlessly than before, working as a collaborator alongside the user.

Perplexity Comet browser: Perplexity launched its AI-native web browser called Comet in September of this year. The browser is built on the Chromium framework, but it leverages autonomous agentic AI to perform tasks on a user's behalf. Comet has the ability to browse sessions, navigate websites, fill out forms, and also conduct online transactions.

OpenAI Atlas (AI-browser): To counter Comet and Google’s AI Mode, OpenAI also introduced its AI-powered browser called Atlas. It is an interactive and agentic AI-based browser that can perform multi-step actions, online shopping, or deep research on the user’s behalf. Its “Memory” feature is what makes the browser more personalised, as users can circle back to previous searches with just a single prompt.

Sora 2 text-to-video generation: Sora 2 is OpenAI’s second-generation video and audio model, which gained huge popularity during its launch. This year, the company introduced several new features, such as Cameos, Remix culture, and a social media-like AI-native Feed, where users can share AI-generated videos. While it's only available in the US and Canada, we expect a broader rollout next year. 

OpenAI’s ChatGPT 5.2: Earlier this year, OpenAI launched its new generation flagship AI model, GPT-5, but in December 2025, it rolled out an upgraded version dubbed GPT 5.2 with capabilities human-expert reasoning, improved visual interpretations, and more. In addition, the AI model is optimised for agentic workflows. This model competes with Google’s latest Gemini 3 model. 

Adobe Max AI suite: Lastly, Adobe made some crucial AI announcements for the year, starting with the new Firefly Image Model 5, for AI images. The company also introduced its video model and AI tools like Generate Soundtrack, Generate Speech, Firefly Video Editor, and more. In addition, the company also introduced its conversational AI assistant, Photoshop Express AI Assistant, that can perform edits on the user’s behalf.