TGViewer
HN Best Comments HN Best Comments @hn_best_comments · 4.16K subscribers
Post #33931 450
Re: Mistral Large 4

Surprisingly it only supports reasoning "none" or reasoning "high".

That setting didn't seem to make any real difference - it added a tiny bit of thinking trace and high actually produced less output tokens than none.

The high bicycle frame is better then the none one though.

Pelicans: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...

(Definitely the best I've seen from any Mistral model: https://simonwillison.net/tags/pelican-riding-a-bicycle+mist... )

simonw, 15 hours ago
More from @hn_best_comments
  1. Oct 7, 2026Re: Paramount Skydance has completed its $111B merger with Warner Bros. Discovery This is…
  2. Oct 7, 2026Re: Sharing AI progress in mathematics > We prove the Unique Games Conjecture The Unique G…
  3. Oct 7, 2026Re: Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates Last n…
  4. Oct 7, 2026Re: Sharing AI progress in mathematics > Several hours of work with Fable simply convinced…
  5. Oct 7, 2026Re: Tell HN: GitHub refuses to remove cracked copies of my software after a month First of…
  6. Oct 7, 2026Re: Sharing AI progress in mathematics I'm still digesting the proof and translating a bit…
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →