cd /news/large-language-models/sanity-check-my-qwen3-8-5090-results… · home topics large-language-models article
[ARTICLE · art-108567] src=forum.level1techs.com ↗ pub= topic=large-language-models verified=true sentiment=· neutral

Sanity check my Qwen3.8 5090 Results; new to Local AI

A user running Qwen3.8-27B on an RTX 5090 with the ninfer runtime reported a 1.72× speedup over LM Studio/llama.cpp, achieving 124.08 tok/s decode and 8,143.82 tok/s prompt eval with mixed NVFP4/FP8 quantization, versus 72.03 tok/s and 2,430.18 tok/s with Q6_K. The user noted that token speeds vary by output type and suggested comparing identical prompts for accurate benchmarking.

read1 min views3 publishedAug 24, 2026

Here’s a datapoint from my 5090; I’ve been running qwen3.8 with GitHub - Neroued/ninfer: High-performance single-GPU inference for selected model checkpoints and GPUs. · GitHub . Similar to you some of this local model stuff has outpaced me since I last checked in, don’t completely grasp the impact of MTP and KVcache quant yet. But this ninfer seemed to have some optimizations for single gpu 5090 arch, but it may not work with the finetunes you need. I also pulled one of the uncensored models with lmstudio ran with llama-benchy as a comparison.

specs

9950x3d

5090

96GB 6400mhz

linux

Model / Runtime Quant MTP Max Decode Prompt Eval Approx. VRAM Relative Speed
JonathanColetti/Qwen3.8-27B-Uncensored-GGUF via LM Studio/llama.cpp Q6_K 2 72.03 ± 5.66 tok/s 2,430.18 ± 24.45 tok/s ~28.5 GB total GPU use 1.00×
neroued/Qwen3.8-27B-nvfp4-NInfer via NInfer mixed NVFP4/FP8 3 124.08 ± 10.56 tok/s 8,143.82 ± 52.68 tok/s ~28.5 GB total GPU use 1.72×

Looks like these models tok/s can vary pretty heavily based on the type of output being generated. So, hard to say how useful this will be for you without comparing identical prompts. I’m interested in your usecase though, uncensored model + red teaming on 5090. Let me know if you find some optimal config with vLLM

── more in #large-language-models 4 stories · sorted by recency
── more on @qwen3.8-27b 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/sanity-check-my-qwen…] indexed:0 read:1min 2026-08-24 ·