DeepSeek V4.1 Flash 1017 (Max)

Re:DeepSeek V4.1 Flash 1017 (Max) by drwiggly on Monday October 19, 2026 @11:02AM (#50153345)

I ran the Strata project with this on 32gb ram 16gb AMD 7900XTX. It goes about 20tok/s off the ssd. I'm guessing 64gb would go to 40tok/s. https://github.com/Niko1221/Strata Quality wise its not fast enough for me to eval. I'll probably spend some tokens on it through openrouter. DeepSeek 4 flash though is a mess compared to GPT Sol 6.1.

[ Parent | on HN ]

Re:DeepSeek V4.1 Flash 1017 (Max) by 72deluxe on Wednesday October 21, 2026 @05:56AM (#50178271)

Thanks (and everyone else) for the reply. I will stick with llama.cpp running llama-server Qwen3-Coder-30B-A3B-Instruct-NVFP4 for now.

[ Parent | on HN ]