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.

cerebras: 450 tokens/sec llama 3.1 70B

7 pointsby davidfiala9 months ago

2 comments

IronWolve9 months ago
Cerebras fails the &quot;how many r&#x27;s in strawberry&quot; test. Grok is the only one who passed that test.<p>Going to be interesting to see the speed and accuracy keep increasing, cant imagine how fast&#x2F;accurate things will be in a decade. Cant wait.
davidfiala9 months ago
- 1,800tps on llama 3.1 8B<p>- 450tps on llama 3.1 70B<p>free chat interface is at: <a href="https:&#x2F;&#x2F;inference.cerebras.ai" rel="nofollow">https:&#x2F;&#x2F;inference.cerebras.ai</a> (requires login)