0546 / 2454

qwenfast — fastest Qwen3.8-27B inference

About: qwenfast is a high-throughput Qwen3.8-27B inference stack using speculative decoding and a custom Triton DeltaNet kernel, hitting 133 t/s on H200.
Product: qwenfast
Company: qwenfast
Category: Developer tools
Posted: Aug 28, 2026
Duration: 0:29
Views on X: 21,389snapshot Sep 14, 2026
Author: @HarshalsinghCN
Original: View on X
Tags: qwen · inference · open-source

Explore similar launches to qwenfast

Poster for EvilCharts — ECharts rendering engine

EvilCharts — ECharts rendering engine

EvilChartsDeveloper tools

Poster for Inco Splash — open-source inference on Apple silicon

Inco Splash — open-source inference on Apple silicon

IncoDeveloper tools

Poster for ThreeJS Super Terrain — UE5-inspired mesh terrain

ThreeJS Super Terrain — UE5-inspired mesh terrain

vibe-stackDeveloper tools

Poster for Omnigent — an open source AI meta-harness for agent workflows

Omnigent — an open source AI meta-harness for agent workflows

DatabricksDeveloper tools