Qwen3.8-27B prefills 2.5x faster on my DGX Spark with TensorFold 1.2.2: 354 to 889 tok/s…
This is a AI post classified by AGK Intelligence as Open & local models (a model release), kept by the AI Radar because it carries real work, not commentary.
Qwen3.8-27B prefills 2.5x faster on my DGX Spark with TensorFold 1.2.2: 354 to 889 tok/s at 2K 1 DGX Spark, 128 GB, TensorFold 1.2.1 Prefill at 2K: 354 tok/s Prefill at 128K: 311 tok/s First token at 32K: 96.3 s Decode, code: 45.9 tok/s Decode, prose: 42.9 tok/s 500 words (code and prose): about 15.8 s 8 requests at once: 60.9 tok/s in total Same Spark, TensorFold 1.2.2 Prefill at 2K: 889 tok/s Prefill at 128K: 660 tok/s First token at 32K: 40.5 s Decode, code: 49.0 tok/s Decode, prose: 45.6 tok/s 500 words (code and prose): about 15.3 s 8 requests at once: 108.8 tok/s in total Qwen3.8-27B
Posted by Niklas (2.8k followers) 1 days ago · 16 likes · 1.5k views · view the original post on X. Kept by the AI Radar as Open & local models. Tools mentioned: niklaslenz.com.
More AI work like this
- Qwen 4 is in my hands and I can't stop grinning 🔥 — @aijoey
- Qwen 4 (unnamed quant) — @loktar00
- 北冥有鱼，其名为鲲。 — @NFT_Chen
- If you want to get started with Local AI but dunno where to start, THIS IS FOR YOU 🚨 — @javierbits
- 💥重磅！Qwen3.8-Flash-Next 量化版炸裂升级：Token消耗直接腰斩50%，准确率暴涨15%！本地32GB丝滑跑通125B级MoE！ — @NFT_Chen
- This could be so cool. — @niklaslenz_ai
- Qwen is teasing a new model — @RoundtableSpace
- 时代变了，以前是token每秒，有人提出了秒每token。。。。 — @mylifcc
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 65.5k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 16:50 UTC. Full method.