This week has been nothing short of groundbreaking in the world of artificial intelligence. From OpenAI’s game-changing features to Google’s free AI models and innovative app-building tools, the AI landscape is evolving at an unprecedented pace. Here’s a detailed roundup of the most exciting developments in AI this week, based on the latest updates.
1. OpenAI’s 03 Mini and Deep Research
OpenAI has launched the 03 Mini model, a highly capable AI designed to tackle everything from intricate math problems to advanced PhD-level science queries. The standout feature? It’s completely free on the ChatGPT plan. Users can pair 03 Mini with OpenAI’s search functionality, enabling the AI to process information and browse the web simultaneously for even more comprehensive results.
During testing, 03 Mini demonstrated its impressive speed by creating a fully functional one-page website in just 4 seconds, complete with working links and a polished layout. The more advanced version, 03 Mini High, took slightly longer (17 seconds) but delivered comparable outcomes. However, both versions faced challenges in coding tests, particularly when tasked with building a fully functional space invader game, encountering issues like distorted backgrounds and laggy gameplay.
Despite these limitations, 03 Mini showcases OpenAI’s commitment to making advanced AI tools accessible and versatile for a wide range of users.
The standout innovation, however, comes from OpenAI’s Deep Research feature, accessible through the $200 Pro Plan. This powerful tool can create detailed 10,000-word reports in a matter of minutes, complete with comparison charts, source references, and clear, easy-to-understand breakdowns of complex subjects. By transforming intricate information into actionable insights, Deep Research is revolutionizing the way professionals and researchers work.
Though currently restricted to three reports per day, this feature is undeniably transformative, setting a new standard for efficiency and depth in AI-driven research.
2. Google’s Gemini 2.0 and Flash Thinking
Google is making headlines with its Gemini 2.0 models, particularly the Flash Thinking with Apps feature. This free, user-friendly tool allows seamless integration with Google’s ecosystem, enabling users to analyze YouTube videos, build apps directly from their phones, and more.
In a live demonstration, the AI showcased its efficiency by creating a self-playing snake game in seconds and providing actionable insights from website analytics. For example, it identified that a website with 19,400 clicks and 73,000 impressions over 16 months had a low click-through rate (2.7%) and an average search position of 13, highlighting opportunities for optimization.
By offering these advanced AI tools for free, Google is setting itself apart in a market where competitors often charge premium prices for similar features. Additionally, developers can use these tools to create custom AI applications, benefiting both casual users and professionals. This strategic move reinforces Google’s commitment to making cutting-edge AI technology accessible to all.
3. GitHub Copilot’s Agent Mode
GitHub Copilot is leveling up. The AI coding assistant, known for suggesting code snippets, is now taking charge with agent mode, a new feature that turns it into an active participant in software development. Instead of just offering suggestions, Copilot can now break down tasks, write code, and fix errors autonomously—while humans oversee the process as managers rather than hands-on coders.
Since its 2021 launch, Copilot has become a staple for developers, especially in VSCode. But competition is heating up. Tools like Cursor, Replit Agent, and Windsurf are pushing boundaries, offering everything from standalone IDEs to fully autonomous coding agents.
GitHub isn’t stopping at agent mode. They’ve also teased Project Padawan, an AI agent designed to automate bug fixes in code repositories. Padawan writes, tests, and refines solutions, with humans stepping in only to review and approve.
This shift raises big questions: As AI takes over coding, will developers become AI managers? One thing’s clear—software engineering is evolving fast, and GitHub is leading the charge.
4. Pika launches ‘Pikadditions’
Pika, the AI-powered social video platform, has rolled out its newest feature, “Pikadditions,” adding even more creative tools to its growing arsenal. This launch comes hot on the heels of the 2.1 model upgrade, which brought 1080p high-resolution videos, enhanced character control, and cinematic camera movements to the platform.
Pika’s latest feature, Pikadditions, introduces a video-to-video model that lets users seamlessly integrate any person or object into their existing videos. Whether it’s spicing up an Instagram Reel or reimagining a classic movie clip, creators can now add surreal, comedic, or unique elements to captivate their audiences like never before.
According to Demi Guo, Pika’s Founder and CEO, “Our goal with Pikadditions is to empower users to push the boundaries of their creativity. We believe that by making advanced AI tools user-friendly and accessible, we can inspire a new wave of content creation that is both innovative and fun.”