MyColo measured local AI inference

Share card

bundle 7 · measured result · 70B+

2.5 tok/s

hardware
NVIDIA GeForce RTX 5090
model
qwen2.5-72b · 424bad2cc13f72f5…
engine / quant / ctx
ollama · Q4_K_M · ctx 4096
workload
wl-big-qwen25-72b-ollama
trust
measured · lone · test · 2026-09-20 21:45:16

Throughput comparison

2.5 tok/s · baseline 2.7 · -5.8% · n=10

70B+ · qwen2.5-72b · your prior signed/measured runs (same recipe)

Same handle, hardware class, size track, and recipe version. Current run excluded.

hash
969829dfad391c153057319e8f99fc579691327d913db5a13f308143c7376012

Hardware class

{
  "cpu": "AMD",
  "driver": "580.126.18",
  "gpu_or_soc": "NVIDIA GeForce RTX 5090",
  "os_name": "Linux",
  "stack": "cuda",
  "vram_or_unified_mib": 32607
}

Sign-offs

hardware receipt · JSON · measure