llamaperf

Qwen3.8 125B (6B active) Flash-Next

on NVIDIA RTX 4070 Super · Strata · 256,000 ctx

Tone: mixedNVIDIA hardware
Oct 14, 2026
Throughput
30-38 t/s gen · 950-1800 tokens/s prompt processing (prefill)Prompt processing measures input prefill speed; it does not tell you the time to first token.
System RAM
32 GB
VRAM reported
12 GB

Summary

User reports Qwen3.8 Flash-Next running at 30-38 tok/s output and 950-1800 tok/s prefill on an RTX 4070 Super 12GB with 32GB system RAM. Setup is Strata with 256K context, MTP and vision off, model partially offloaded to system RAM. User notes speeds depend heavily on CPU, SSD, RAM speed and GPU bandwidth, and that 16GB VRAM may allow MTP.