I’m launching AI IQ

A new project tracking frontier AI models on the human IQ scale.

I’m launching a new project today: AI IQ.

The idea is simple: frontier AI models are getting too important to understand through scattered benchmark tables, vibes, and launch-day hype.

So AI IQ tracks the leading models — GPT-5.5, Claude Opus 4.7, Gemini 3.1, Grok 4.3, Kimi K2.6, Qwen3.6, DeepSeek V4, Muse Spark, and others — and scores them on a more intuitive scale: the human IQ scale.

Instead of asking whether a model got 72.4% on one benchmark and 81.9% on another, AI IQ asks:

• Where would this model land on the IQ bell curve?

• How quickly is frontier intelligence improving?

• Which models are strongest on IQ, EQ, and agentic capability?

• What does intelligence cost in practice?

• Which models are actually worth using?

I’m starting with four posts and a live site:

https://aiiq.org/

The newsletter will live here:

I’m keeping this separate from my personal Substack because AI IQ is one of a few products I’m building, and I want this newsletter to stay useful as a place where I can share launches, updates, and broader thoughts.

But if you’re interested in AI models, benchmarks, software, investing, or the question of how machine intelligence is actually progressing, I’d love for you to subscribe to AI IQ.

The goal is to make model progress legible.

Not just “which model is #1 this week,” but what is actually changing, what matters, what is noise, and how close these systems are getting to different kinds of human capability.

Curious which chart surprises you most.