BuildBrief: GPT-5 vs Claude 4

OpenAI has replaced all their models: GPT-4.1, GPT-4o, o3, o4...
👉 with one new model: GPT-5.

It's meant to be faster, smarter, and easier to use.

But is it actually better than Claude 4?

Here are the big things:

(1) No more model-picking

You don't choose the AI model for specific tasks anymore... GPT-5 decides for you.

If you've got a simple question - it answers instantly.
If you've got a complex problem - it slows down to think it through.

(2) Closest to AGI right now

GPT-5 feels like the most "general" AI we've had yet.

It's way better at planning, coding, and can hand you a working app from a single prompt... even making decisions without you asking.

It speeds through simple tasks, slows down for complex ones, and spots issues most models miss.

Example: when I asked GPT-5 to embed videos to my blog, it caught a security risk whilst building it - which Claude 4 didn't even notice.

HONEST TAKE:

GPT-5 is impressive, but it's slow when coding... especially in Cursor.

So for now I'm sticking with Claude 4 since they're basically just as good for building, but Claude 4 is much faster.

---

Happy building!
– Kristoffer, BuildWithAI