TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

DeepSeek-V3-0324

5 pointsby desideratumabout 2 months ago

1 comment

reissbakerabout 2 months ago
It&#x27;s &quot;just&quot; a minor version update, but in my own testing it seems <i>much</i> stronger than the original V3 — basically on par with R1 for the usual tricks I throw at LLMs, without needing &lt;think&gt; tokens.<p>I&#x27;m sure they&#x27;re re-RL-training an R1-[minor bump] on top of this model, or perhaps even an R2; it&#x27;ll be extremely strong when it comes out. For now I&#x27;ve swapped most of my usage to this new V3, since it&#x27;s basically on-par for my use cases with R1 and doesn&#x27;t require waiting for thinking tokens.