←back to thread

Llama 3.1 405B now runs at 969 tokens/s on Cerebras Inference

(cerebras.ai)
426 points benchmarkist | 2 comments | 19 Nov 24 00:15 UTC | HN request time: 0.403s | source
Show context
bargle0 ◴[19 Nov 24 03:49 UTC] No.42179943[source]▶
Their hardware is cool and bizarre. It has to be seen in person to be believed. It reminds me of the old days when supercomputers were weird.
replies(1): >>42179953 #
1. IAmNotACellist ◴[19 Nov 24 03:51 UTC] No.42179953[source]▶
Don't leave us hanging, show us a weird computer!
replies(1): >>42181092 #
2. campers ◴[19 Nov 24 08:21 UTC] No.42181092[source]▶
https://web.archive.org/web/20230812020202/https://www.youtu...