llamaperf

Qwen3.8 27B

on NVIDIA RTX 4090 · 132,000 ctx

Tone: positiveNVIDIA hardware
Oct 7, 2026
Throughput
125.0 t/s gen

Use cases

vision

Summary

User reports Qwen 3.8 27B with vision and MTP at 125 tok/s on an RTX 4090. Setup uses the integrated GPU for display output, freeing about 2.5GB of VRAM, which raised the context window from 65k to 132k. The user notes swapping the display cable back for gaming is a minor annoyance.