Qwen3.8-27B hits 223 tok/s on RTX 6000 Pro with NVFP4
BanghuaZ · x · 2026-08-17
This repository provides scripts to run Qwen3.8-27B via SGLang on NVIDIA RTX 6000 Pro (96GB). Utilizing NVFP4 weights and DSpark speculative decoding, it achieves 200-223 tok/s single-stream throughput. The setup supports the native 256k context window and 8 concurrent requests by default, using FP8 KV cache for optimal memory usage.
Related event: Qwen3.8-27B local benchmarks: from RTX 4080 Super to RTX 6000 Pro(7 posts)→
More from Infra
- Musk says he wants 'maybe more than 10GW' of compute; Nvidia now counts SpaceX as a CSP and co-designs a space server — elonmusk · 2026-10-16
- Jensen Huang: Musk's AI buildout moved from K's to M's, now a CSP-tier NVIDIA customer — elonmusk · 2026-10-16
- EpochAI pegs GPT-6 training at up to 1.8e27 FLOP; scaling01 calls param estimates way too high — scaling01 · 2026-10-16
- Meta's Muse Spark Open scores 39 vs 48 for closed Spark 1.3 on AA Intelligence Index — EAccelerate_42 · 2026-10-16
- LLM spots useless 95GB-per-frame memcpy in ffmpeg dblur, cutting runtime from 2452ms to 49ms — IgorCarron · 2026-10-16
- DeepSeek-V4-Flash on 4x Radeon AI PRO R9700: the 'refused' 4-card config hits ~79 tok/s — brainchillzZ · 2026-10-16