Skip to content
llm-speed
wrapped·Apr 28, 2026
/r/r_h0-use1ypnb
your run

Llama-3.1-8B-Instruct-4bit

29.2
tok/sdecode
M3 Pro (18-core GPU) + 36GB unified
rank in tier
4/10M3 Pro runstop 40%
best workload
chat-short
where the rig flew
slowest workload
—
single-workload run
backend
mlxMLX is the fastest backend on Apple Silicon for dense models.
faster than
  • Qwen3.8-27B-4bit on M3 Ultra27.3 tok/s
  • Muse-Spark-Open-816B-A42B-UD-00001-of-00005 on 3x A10026.3 tok/s
  • llama3.2 on AMD64 Family 23 Model 104 Stepping 1, AuthenticAMD23.9 tok/s
share
29.2 tok/s on M3 Pro running Llama-3.1-8B-Instruct-4bit — top 40% of M3 Pro runs. Signed benchmark via llm-speed.
https://llm-speed.com/wrapped/r_h0-use1ypnb

This is a shareable view of one signed llm-speed suite-suite-v1 run. Numbers are not edited. The raw record has every workload, the public key, and the fingerprint hash.