# Opus 5.5 max over-thinks; GPT-6 Luna is worth a try

> A quick take · Simon Willison, September 22, 2026

---

LLMS index: [llms.txt](/llms.txt)

---

> Claude Opus 5.5 max over-thinks to the point of breaking
>
> This makes me suspect that “max” is effectively useless—if it over-thinks to breaking point on a stupid SVG prompt I don’t trust it not to do the same for more interesting work.
>
> — [Simon Willison · Opus and Sol and Luna](https://simonwillison.net/2026/Sep/22/opus-and-sol-and-luna/)

Max really isn't cost-effective unless the model can avoid over-thinking on its own.

> GPT-5.6 Luna was already my favorite model for building applications against, because it combined excellent performance with being really cheap.
>
> — [Simon Willison · Opus and Sol and Luna](https://simonwillison.net/2026/Sep/22/opus-and-sol-and-luna/)

GPT-6 Luna is worth a try.