Qwen3.8-Flash-Next: EXL3 ~600 vs NVFP4 ~150 tok/s reported concurrent decode. 🤯
This is a dev post classified by Jev as Other (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
Qwen3.8-Flash-Next: EXL3 ~600 vs NVFP4 ~150 tok/s reported concurrent decode. 🤯 Both on DGX Spark GB10. The interesting part is what the engines actually do with those compressed weights. This will be my last informational post on Qwen3.8-Flash-Next as im cooking up something new for you all! 👨🏻🍳 3.05 bpw EXL3 through my ExLlamaV3 fork. NVFP4 through vLLM. 𝗧𝗛𝗘 𝗗𝗘𝗖𝗢𝗗𝗘 𝗗𝗜𝗙𝗙𝗘𝗥𝗘𝗡𝗖𝗘 Balanced workload, EXL3 / NVFP4: Single-request decode: 78.3 / 31.0 tok/s Per-request decode at 8 requests in flight: 75.9 / 18.9 tok/s Time per output token at that load: 13.2 / 53.4 ms Reported concurrent dec
Posted by Cruz (2.1k followers) 5 h ago · 49 likes · 2.7k views · view the original post on X. Kept by the Dev Radar as Other. Tools mentioned: qwen3.8-flash-next-exl3-dgx-spark-recipe, exllamav3.
More dev work like this
- Dev Ojha (@zkDragon) explains how Zakura is rebuilding Zcash infrastructure for scale: — @arjunkhemani
- Anthropic’s new Mythos class model SVG — @HarshithLucky3
- 1.0-release了，增加了mqtt控制。利用小爱音响系统内自带的MQTT… — @eastwoodnet
- 🔥腾讯混元 Hy-Image 3.5 Preview 正式上线！内部几百名设计师 GSB 盲测，直接对标一线生图模型！ — @NFT_Chen
- 很有意思的图啊，MiMo 是想一步一个脚印靠自己爬～ — @Fei2411
- Maps the world's infrastructure using data from OpenStreetMap. — @tom_doerr
- emergency services? … yes, I’d like to report a chart crime 🚔 — @theaaron
- So you like Bonsai, — @Mr_WestCoast
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 20.7k posts from 4.9k X accounts over the last 21 days, 2.4k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 08:21 UTC. Full method.