Comment by onion2k

21 hours ago

I just checked gpt-oss:20b on my M4 Pro 24GB, and got 400.67 tokens/s on input and 46.53 tokens/s on output. That's for a tiny context of 72 tokens.

0 comments