alt Hacker NewsDeepSeek 4.1 Flash 0910 is perfect for M5 Ultra 256GiB. Running it fully resident in RAM, prefill at ~2500 tok/s and decode at ~40 tok/s. Probably tons of room to improve from there.
Replies
sdg03ksdv0d • today at 8:27 PM
You are running this now? :o
Which quantization are you using there?