llamaperf
Calculator
GPUs
Engines
Guides
Blog
Community
MCP
Advertise
Submit
Sign in
Submit
llamaperf
/
M5 16GB
/
report
Qwen3.5 35B (3B active)
on
M5 16GB
Apple hardware
Sep 29, 2026
Share
X
Reddit
LinkedIn
Bluesky
Copy
link
Throughput
17.0 t/s gen
System RAM
16 GB
Summary
User reports Qwen3.5 35B A3B running at roughly 17 t/s on a MacBook Pro M5 16GB. Setup uses slotstream; no engine, quant or context length stated. The figure is given as approximate.
View source ↗
Feedback