Elon Musk just dropped a $300/month AI—and it's not ChatGPT.
It’s called 👉Grok 4.
▶️ It claims to beat PhDs in every subject.
▶️ It can reason better than most humans.
▶️ And it just launched with super-agent mode.

But here's the part that got me:

Grok 4 isn’t just another language model.
According to Musk, it’s post-grad level in every subject.
And the premium tier—SuperGrok Heavy—spawns multiple AIs to work on a single problem like a study group. Then they vote on the best answer.

It even beat OpenAI’s latest on the “Humanity’s Last Exam.”
No tools. Just raw reasoning.
📊 25.4% score without tools.
📊 44.4% score with tools (beating Gemini and GPT-4o variants).

But here’s where it gets complicated.
Last week, Grok’s official account on X (formerly Twitter) made antisemitic comments.

It sparked major backlash—and Musk’s team tried to quietly tweak Grok’s system prompt.

That’s the tension right now:
▶️A model that pushes benchmarks forward
▶️A team that’s still playing cleanup on misalignment

As someone building in AI, this launch gave me whiplash.
▶️On one hand, the tech is wild.
▶️On the other, it reminds me how brittle alignment still is.

Would you pay $300/month for this?
Is xAI bold—or just reckless?
👇 Drop your take in the comments.
🔗 Full breakdown + newsletter here: https://www.npmix.com/blog/grok-4-is-here-and-its-equal-parts-genius-and-nightmare