Re:Local Qwen isn't a worse Opus, it's a different tool
by te0006 on Friday June 19, 2026 @09:45AM
(#48598518)
Interesting setup. What GPU(s)/VRAM, CPU and RAM are you using for the 122B model, with which quantization, and what token rates do you achieve for prefill and generation?
[ Parent
| on HN ]
There are no comments on this comment yet.