LINKED LIST [txt mode] ▸ antirez/ds4: DeepSeek 4 Flash and PRO...
home explore | log in

antirez/ds4: DeepSeek 4 Flash and PRO local inference engine for Metal, CUDA and ROCm

github.com · 2026-09-25 · 1 upvotes

log in to save, upvote or flag this.


─── In 0 lists ─────────────────────────────────────────

(not in any lists yet)

* prabanta-dev/janas: 37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU. (from the discussion)
* Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash? (from the discussion)
* From the creator of Redis; run LLM locally with ds4 (from the discussion)
* Local AI Needs to be the Norm (from the discussion)

─── Discussions ────────────────────────────────────────

* DeepSeek 4 Flash local inference engine for Metal
178 pts · 56 comments · node