Back to AIBriefs
AnalysisAI Models

Qwen3.6 27B achieves 1000 t/s with V100s in concurrent setup

Qwen3.6 27B achieves 1000 t/s with V100s in concurrent setupView on Reddit

The setup achieved 1000 t/s generation on Qwen3.6 27B with 128 concurrent requests using V100 GPUs. For single user, generation speed drops to 80 t/s with 3000 t/s processing.

·
23 days ago
Qwen3.6 27B achieves 1000 t/s with V100s in concurrent setup — AIBriefs