llamaperf
Calculator
GPUs
Engines
Guides
Blog
Community
MCP
Advertise
Submit
Sign in
Submit
llamaperf
/
GPU
/
report
Qwen3.8 27B
on Unknown GPU · Inco Splash
Sep 29, 2026
Share
X
Reddit
LinkedIn
Bluesky
Copy
link
Throughput
74.0 t/s gen
Summary
User reports Qwen3.8-27B at 74 tok/s on a Mac. Setup uses the Inco Splash engine. The post title gives no Mac chip or memory size, so the hardware is not identified.
View source ↗
Feedback