🧠 This is it! — #9 The First AIs With 130+ IQs Started Popping Up. Seriously!

Not long ago, language models were clever assistants — parroting back facts, drafting emails, maybe summarizing PDFs. But now?

They’re acing IQ tests. Real ones. Designed for humans.

https://trackingai.org/home

Non-members can read here or subscribe to stay updated whenever I publish.

Over a year ago, I wrote a quick post titled: “This is it! #1 — Turks are finally putting AI to proper use”. It was a lighthearted take on local creativity with LLMs. What followed was unexpected — a steady stream of “This is it!” moments from around the world: AI therapists in Turkish hospitals, GPT-powered supply chain bots in Africa, AI musicians, voice actors, farmers, coders, risk analysts…

More than 30 posts later, and 400,000+ impressions across platforms, the series has become a kind of living documentary of how AI is colliding with everyday life. And this post?

Feels like a chapter break. The kind you remember.

🧠 AIs With 120+ IQs Are Here. And One Just Hit 136

Not just passing — scoring 120+, placing them above 90% of the global population. Yes, seriously.

In the latest Norway Mensa IQ evaluations, Grok4 model obliterated the competition. It didn’t just edge ahead — it left others in the dust, with scores significantly higher than GPT-4, Claude, Gemini, and the rest.

What this means? AI is no longer just “useful.” It’s becoming brilliant.

📈 Let’s Put It in Perspective

• I’m from Turkiye and the national IQ average there is: ~94
• Western Marmara (Most developed, highest income region): ~98
• OpenAI o3 Pro vision: ~96
• Grok 4: ~136

That’s a 38-42 point leap — the equivalent of jumping from “average” to “superior reasoning ability.” Or more concretely:

If GPT-4 is your helpful classmate, and this was 2024… Grok 4 is that freakishly sharp kid who finished the test 10 minutes early and aced it in 2025…

🤔 Why Does It Matter?

Because it’s not just about fun with numbers. IQ tests are proxies for:

• 🧩 Pattern recognition
• ⚙️ Learning speed
• 🔍 Abstract reasoning
• 🔧 Problem-solving

When an AI performs at 130+ on a test designed to evaluate human intelligence, we’re not just dealing with a smarter search engine — we’re looking at a reasoning agent.

That changes everything.

🛠️ Use Cases That Just Got Supercharged

If you have an AI with a 130+ IQ, here’s what you can ask it to do (and expect results):

• Solve puzzles that you couldn’t crack
• Analyze image + text together (vision + logic, a major frontier)
• Beat your entire fantasy football league 😅

More importantly (this is where you start losing your job);

• Digest legal and financial documents and spot contradictions
• Design multi-step strategies (marketing plans, startup roadmaps, even negotiation tactics)
• Understand nuance, context, and strategic intent

🧠 But Wait… It’s Still Just a Preview?

Yup. The version of LLMs that hit these scores aren’t even final. It’s not clear how it was fine-tuned, how “test-aware” it was, or what tricks were used in reasoning chains — but what’s undeniable is this:

AI just hit a milestone in raw cognitive ability. And no one was really ready for it…

🔮 Where Do We Go From Here?

If you thought AI was just about chatbots and productivity hacks, think again. We’ve just crossed a line:

Artificial intelligence now outperforms most humans on general reasoning tasks.

It doesn’t mean AI is conscious. It doesn’t mean it’s infallible. But it does mean one thing very clearly:

We’re no longer the smartest species on the test sheet.

This is it. Seriously.

📌 *Subscribe to stay updated with future “This is it!” drops. Real-world AI breakthroughs. No hype. Just signal.*

#AI #IQ #LLM #OpenAI #o1 #Breakthrough #ThisIsIt

All the best

Altan

P.s.

For the previous post in this series ->

[embed]🐄This is it! #8 — This Time AI in the Barn Not in Silicon Valley. Not in a Cambridge lab. But in Karacabey, Bursa, at a family-run dairy farm called Feyz…medium.com