DeepSeek V3 is now the highest scoring non-reasoning model | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		DeepSeek V3 is now the highest scoring non-reasoning model (twitter.com/artificialanlys)
		14 points by aurareturn on March 25, 2025 \| hide \| past \| favorite \| 3 comments

buyucu on March 25, 2025 [–]

> This is the first time an open weights model is the leading non-reasoning model, a milestone for open source.

wow, this is huge!

ai-christianson on March 25, 2025 | [–]

It's performing well on the aider benchmark too.

This one is legit.

buyucu on March 25, 2025 | | [–]

The original Deepseek V3 model was amazing. If this one is better, then I can't even imagine that.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact