- AI Adopters
- Posts
- OpenAI launches o3 and o4-mini
Welcome back.
OpenAI just rolled out o3 and o4-mini, its most capable reasoning models yet.
With smarter tool use, faster performance, and stronger visual understanding, they’re built to power the next generation of AI assistants.
Let’s dive in.
In today’s release:
1. OpenAI launches o3 and o4-mini
2. Google gives students AI tools
3. Microsoft’s “computer use”
OpenAI o3 and o4-mini push agentic AI forward
OpenAI has launched o3 and o4-mini, its most advanced reasoning models to date.
o3 excels at coding, math, and scientific reasoning, while o4-mini delivers strong performance at lower cost and latency. Both models use a refined approach to multimodal problem-solving, allowing them to interpret and reason with images, data, and documents in more human-like ways.
Full tool access: Models can browse, analyze files, use Python, and generate images
Visual reasoning: Interpret whiteboards, sketches, and charts, even blurry or rotated ones
Performance gains: o3 reduces major errors by 20% vs. o1 in expert evaluations
o4-mini speed: Fast, cost-efficient reasoning with top scores on AIME 2024 and 2025
The models are built on a new reinforcement learning scale-up, allowing them to think longer when needed without sacrificing speed. Tool use is more strategic, enabling models to pivot mid-task, run multiple web searches, or chain steps together in ways that feel purpose-driven.
How can you use this to your advantage?
If you're working with messy data, complex workflows, or real-world research, o3 and o4-mini let you move beyond static Q&A and start building assistants that reason, adapt, and solve. They're ideal for anyone creating tools that require deeper thinking and reliable multi-step execution.
Google gives college students free AI tools
Google is giving U.S. college students access to its most powerful AI tools completely free through spring 2026.
Students who sign up by June 30, 2025, will unlock AI tools for writing, studying, research, and even video generation. Features like Deep Research, Canvas, Veo 2, and Gemini Live are powered by Google’s top model, Gemini 2.5 Pro.
Full access: Gemini Advanced + NotebookLM Plus included free
Study support: Create audio summaries, mind maps, and guided study plans
Creative tools: Generate videos with Veo 2 and animations with Whisk
Storage boost: 2TB of Google One space for projects, media, and more
Google says the tools are designed to help students do more with less stress, whether it’s analyzing research, visualizing data, or turning notes into podcasts. The offer also includes built-in Gemini access across Docs, Sheets, and Slides, making it a strong academic upgrade for the next two school years.
How can you use this to your advantage?
If you're a U.S. college student, this is a high-value upgrade for your study routine. You’ll get powerful tools to research faster, write better, and stay organized, all without spending a dime.
Microsoft Copilot Studio adds UI automation
Microsoft has introduced a new “computer use” feature in Copilot Studio that lets AI agents control desktop apps and websites through the UI, no API required.
Built on Microsoft’s secure cloud infrastructure, computer use supports major browsers like Edge, Chrome, and Firefox, and requires no local servers. It’s a big step toward more flexible, agentic automation, especially for teams still working with legacy or closed systems.
UI control: Agents can interact directly with software interfaces
No-code: Describe tasks in natural language to build automation
Real-time recovery: Agents adapt to interface changes on the fly
Hosted by Microsoft: No local setup needed, enterprise-grade security included
This release also modernises RPA by making it easier for non-developers to create automations. Makers can view side-by-side reasoning chains, test outputs instantly, and refine automations with natural language.
How can you use this to your advantage?
If your team deals with repetitive tasks across clunky UIs or disconnected tools, Copilot Studio’s computer use helps you build smarter automations without needing dev support. It’s especially useful where APIs fall short.
OTHER AI NEWS
Grok launches Workspaces: You can now organize chats, files, and custom instructions in one place using Workspaces in Grok. It helps you stay focused on specific projects or topics without losing context.
Gemini Live screensharing is now free on Android: Google’s Gemini app now lets all Android users share their camera and screen with the AI assistant at no cost. The feature was originally meant for paid subscribers, but is rolling out for free starting today.
Grok gets a memory feature: xAI has added memory to Grok, allowing it to recall past chats and offer personalized replies. The feature is now in beta in Grok and the mobile apps, with EU/UK support coming later.
Samsung may add Perplexity AI to Galaxy devices: Samsung is in early talks to bring Perplexity’s AI assistant to its devices. If the deal happens, it could boost Perplexity’s visibility on hundreds of millions of devices worldwide.
Docusign unveils AI contract agents: Docusign has launched the AI contract agent, designed to automate tasks and streamline the entire agreement lifecycle. The tool is capable of analyzing contracts in seconds and will roll out by year’s end.
POPULAR AI TOOLS
TheLibrarian.io → Your WhatsApp AI Assistant.
Playbook → Beautiful documentation your customers will love.
Trueguard → Stop fake user signups.
Wiza Monitor → Track job changes and get Slack + email alerts.
Backed AI → Your Daily AI Companion for Back Pain & Posture Correction.
AND THAT’S A WRAP
Thank you for reading!
If you found this email useful, share it with a friend or colleague who also loves AI.
Also, drop me a follow on Twitter/X for more AI and tech updates.
I will talk to you soon!
Mike