llamaperf
Submit

Qwen3.6 27B

on 2× NVIDIA RTX 3090

Tone: positiveNVIDIA hardware
Jun 1, 2026
Share
Throughput
70.0 t/s gen
VRAM reported
24 GB

Use cases

creative-writing

Summary

User reports Qwen3.6-27B at ~70 t/s on dual RTX 3090s, generating interactive HTML content inline with chat.
View source ↗