In partnership with

Welcome, Automaters!

Ever notice how every time a big tech company drops some flashy new AI tool, OpenAI claps back the very next day with something even wilder?

It happened with Google, Anthropic, Mistral, Microsoft. And nowβ€”right after Apple’s WWDC and all those shiny AI announcementsβ€”OpenAI casually rolls out a brand-new reasoning model…that’s kind of making the competition look bad.

So, is this all just a coincidence? Or is OpenAI quietly playing the smartest PR game in tech?

We may never know. But firstβ€”let’s break down what this new model actually does πŸ‘‡

Oh, and speaking of Apple and those AI updates? Yeah, they might actually be a bit of a problem there…..

Here's what we have for you today

πŸ’ƒ OpenAI’s New o3-pro Just Stepped Into the Chat

OpenAI just dropped o3-pro β€” a.k.a. its smartest model yet.

This isn’t just another AI upgrade. o3-pro is a reasoning model, which means instead of blurting out answers like your average chatbot, it actually thinks through stuff β€” step by step. We're talking smarter, more reliable responses, especially in tricky areas like:

  • Math

  • physics

  • Coding

  • Writing help

  • Education

  • Business-y tasks

Basically, if o1-pro was the brainy intern, o3-pro is the PhD-level consultant that’s replacing it.

Unfortunately it’s not for everyone.

As of now, it’s already live for ChatGPT Pro and Team users (as of Tuesday). Enterprise and Edu users get it next week. And for devs, it’s already available in the API.

And about the pricing: It takes $20 per million input tokens and $80 per million output tokens.

Sounds like a lot, but to put it in perspectiveβ€”1 million tokens is about 750,000 words. Basically, this thing can chew through War and Peace and still have room for dessert.

But perhaps the question on your mind is: is it actually better?

Well, yup β€”and not just a little bit. According to OpenAI’s own testing, reviewers consistently picked o3-pro over the original o3 because of its stronger performance in clarity, instruction-following, accuracy, and comprehensiveness.

It also has tool accessβ€”which is a big deal. That means with ChatGPT tools enabled, o3-pro can:

  • Browse the web

  • Analyze files

  • Understand visuals

  • Run Python code

  • Use memory to personalize your chats

That said, it’s not perfect. In fact OpenAI admits that:

  1. o3-pro’s response time takes a little longer than o1-pro’s.

  2. Temporary chats are currently disabled due to a "technical issue."

  3. It can’t generate images.

  4. And it doesn’t support Canvas, (OpenAI’s hyped workspace feature).

But here’s the real mic drop: o3-pro crushed Google’s Gemini 2.5 Pro on a major math benchmark (AIME 2024) and outperformed Claude 4 Opus β€” Anthropic’s top model β€” on a PhD-level science test.

So yeah, o3-pro isn’t just OpenAI’s smartest model β€” it’s their way of saying: β€œWe’re still top dog.”

If you’re doing anything remotely technical, educational, or strategic...this is the model you want in your corner.

10x Your Outbound With Our AI BDR

Imagine your calendar filling with qualified sales meetings, on autopilot. That's Ava's job. She's an AI BDR who automates your entire outbound demand generation.

Ava operates within the Artisan platform, which consolidates every tool you need for outbound:

  • 300M+ High-Quality B2B Prospects, including E-Commerce and Local Business Leads

  • Automated Lead Enrichment With 10+ Data Sources

  • Full Email Deliverability Management

  • Multi-Channel Outreach Across Email & LinkedIn

  • Human-Level Personalization

πŸ€¦β€β™‚οΈApple’s upgraded AI Models Fall Short in Benchmarks

So, Apple finally pulled back the curtain on its upgraded AI modelsβ€”one that runs offline (Apple On-Device) and one that lives in the cloud (Apple Server).

And while it all sounds great on paper, look a little closer and… well, things get kind of humbling.

Let’s break that down:

Apple On-Device runs right on your iPhone or Mac with no internet required. Plus It handles things like:

  • Text generation

  • Summarization

  • Text analysis

  • It’s sleek and compact (about 3 billion parameters), and is now available to developers via Apple’s Foundation Models framework.

While Apple Server, the more powerful model, lives in Apple’s data centers and was supposed to be the real showstopper.

But here’s where things got awkward… 😬

Human testers (anonymous but Apple-approved) ran their own benchmarks and… let’s just say, it didn’t go too well.

  • Apple On-Device was rated β€œcomparable” to similarly sized models from Google and Alibaba.

  • As for Apple Server? Testers actually preferred β€” or to put it bluntly, rated it worse than OpenAI’s GPT-4o (which, btw, dropped over a year ago πŸ‘€).

  • In image analysis tests, Apple’s model got beat by Meta’s Llama 4 Scout… which usually ranks below models from Google, OpenAI, and Anthropic. And the kicker? It’s not even Meta’s best model. Yikes.

But hey, not all is bad news.

  • The new models now support 15 languages

  • Understand image and text inputs

  • Are faster and more efficient

  • Apple also claims improved tool use and efficiency.

Still, the bigger picture’s kinda messy, and here’s why:

  1. It proves that Apple’s clearly lagging behind in the AI race

  2. The much-hyped Siri upgrade has been delayed indefinitely, which only adds to the evidence.

  3. Worst of all, some customers are straight-up suing Apple for promoting AI features that… well, don’t actually exist yet. 😬

In short: Apple’s latest AI update is definitely better than what they had before β€” but when stacked against OpenAI, Google, and Meta?

It’s not giving "industry leader." It’s giving β€œstill catching up… but trying really hard.”

✍️ Editor’s Corner

Hey Automaters,

Sorry. We let you down on the exciting news.

We were supposed to release our new website today. But due to some unforeseen circumstances, we aren’t able to release it just yet.

We were genuinely excited to showcase what we’re working on. We still are.

But it’s taking a little bit longer.

Though I hope that when you do it see it, you’d say that it was worth the wait.

Cheers,

Tak Lo

🧱Around The AI Block

  • πŸ‘¨β€πŸ’» How to build the perfect website using AI (Step-by-Step for Beginners).

  • 😱 News sites are getting crushed by Google’s new AI tools.

  • πŸ€– Mistral releases a pair of AI reasoning models.

  • πŸ” Starling’s AI banking tool shows you how much you’re wasting on McDonald’s.

  • πŸ”Ž Microsoft Edge is getting an AI-powered browser history that works with typos.

  • 🦾 The Trump administration appears to be planning its own chatbot.

  • 🏑 Google just soft-launched nine cool Home app features.

  • πŸ’Έ Meta is reportedly making a $15 billion bet on AGI.

πŸ€– ChatGPT Prompt Of The Day: Generate Auto-Replies That Actually Sound Human

Managing your inbox shouldn't feel like a full-time job β€” especially when you're juggling leads, client work, or growing your business.

This prompt helps you turn ChatGPT into your personal AI email assistant. It doesn’t just reply β€” it replies smartly. With empathy. With context. With tone-matching magic. Whether you're a freelancer or part of a fast-moving sales team, this tool helps you save time and make a solid impression.

Here’s How to Use This Prompt Effectively:

  1. Be Specific with the Industry: The more clearly you define your niche (e.g. β€œindie app developer” vs. β€œtech”), the better the email tone and content will match.

  2. Paste the Full Email Thread if Needed: This gives the AI more context to work with, helping it craft smarter responses.

  3. Define the Tone: Want it friendly, formal, empathetic, persuasive? Say it. That helps fine-tune the reply style.

  4. Review Before Sending: Always skim the reply to ensure it aligns with your voice and intent.

Here’s an example:

You are an AI email assistant specialized in [insert your industry, e.g., "freelance graphic design", "SaaS sales", or "legal consulting"]. Generate customized, professional, and context-aware replies to client or prospect emails. Your replies should match the sender's tone, show empathy or professionalism as needed, and always move the conversation forward.

Here’s a sneak peek:

Upgrade now to see this whole month’s prompt videos and more, or buy TODAY’S WODΒ forΒ justΒ $1.99

Is this your AI Workout of the Week (WoW)? Cast your vote!

Login or Subscribe to participate

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content belowπŸ‘‡

What'd you think of today's email?

Login or Subscribe to participate

Your feedback means a lot to us and helps improve the quality of our newsletter.

logo

πŸš€ Want your daily AI workout?

Premium members get daily video prompts, premium newsletter, an no-ad experience - and more!

πŸ”“ Unlock Full Access

Premium members get::

  • πŸ‘¨πŸ»β€πŸ« A 30% discount on the AI Education Library (a $600 value - and counting!)
  • πŸ“½οΈ Get the daily AI WoD (a $29.99 value!)
  • βœ… Priority help with AI Troubleshooter
  • βœ… Thursday premium newsletter
  • βœ… No ad experience
  • βœ… and more....

Reply

or to participate

More From Capital

No posts found