DeepSeek v4.1 Flash 1017 Quantized to Run on 128GB Macs at 20tps
returnity · reddit · 2026-10-18
Reddit user returnity shipped a PR to antirez's ds4 local inference project enabling the new vision tower of DeepSeek v4.1 Flash 1017, and released a matching GGUF quant on HuggingFace.
The quant follows antirez's original DS4.1 Q2 recipe: 2-bit routed experts with 8-bit+ elsewhere, plus the original FP8 engram table. The wired backbone is just over 150GB and streams from SSD on 128GB machines with routed experts kept in hot cache, hitting up to 20tps decode on an M5 Max; the 200GB engram table is paged in on demand. The revamp changed 40% of routed expert tensors, 30% of engrams and 50% of vision tower experts, and the author calls it the new leading open-weights model.
Related event: Developer adds DeepSeek v4.1 Flash support to antirez's ds4(2 posts)→
More from Infra
- Modal triples valuation in a week while data center owners raise nothing — YvesMulkers · 2026-10-18
- Reddit asks: is 2-bit or 3-bit quantization actually usable for daily LLM work? — Queasy_Back2588 · 2026-10-18
- Samsung and TSMC reportedly team up to produce Nvidia's next-gen HBM chips — Polymarket · 2026-10-18
- Safari Caps Incognito and Low-Power Mode at 30fps — Found via AI-Assisted Bench Bisect — oykun · 2026-10-18
- Running DeepSeek V4.1-Flash (318GB) on a single modded 4090 at 20+ tok/s — lxfater · 2026-10-18
- Hana Securities: YMTC likely to IPO this December, aiming to raise $33B — basedjensen · 2026-10-18