WHAT NEW IN X10.30 FRI

AI collections from X · 22 collections · 60,637 posts

deepseek-v4.1POST

DeepSeek V4.1 Flash on 4x RTX Pro 6000: 250-300 tok/s at c1

Author shares an optimized repo claiming 250-300 tok/s at concurrency 1 and 1000+ tok/s at c8 on four RTX Pro 6000 GPUs, with long context and vision/engram offloading to SSD.

0xSero
@0xSeroArchived 11 days agoOriginally posted 2026.10.18
0xSero @0xSeroOriginally posted 2026.10.18
Open original on X

Calling on 4x RTX Pro 6000 owners Optimised Ds4.1-Flash-1017 repo. I’m getting 250-300 tok/s c1 and 1000+ tok/s at c8 TONS of context and concurrency, vision and engram offloading to ssd, this does p much all my work https://github.com/sybil-solutions/deepseek-v4.1-flash-4x-rtx-pro-6000

x.com/0xsero/status/2111506865972801701Archived 2026.10.18 16:30