deepseek-v4.1IMAGE
550B MoE DeepSeek V4.1 Flash runs at 20 tok/s on dual RTX 3090 desktop
A user runs the 550B MoE DeepSeek V4.1 Flash variant via Project Maya on a 2020 dual-3090 desktop with 128GB RAM and NVMe, hitting 20 tok/s decode and 1,000 tok/s prefill.
Axitek
@axitekArchived 6 days agoOriginally posted 2026.10.25
Axitek @axitekOriginally posted 2026.10.25
We have frontier-level intelligence at home now and I'm not even kidding. DeepSeek-v4.1-Flash-1017:Maya-S, a 550B MOE model, running on a <$5k desktop PC from 2020. 🚀 Decode: 20tps 🔥 Prefill: 1,000tps Thank you Project Maya! https://github.com/mw00/project-maya HW 2x3090+128gb+nvme
x.com/axitek/status/2114153750382428288Archived 2026.10.25 11:59