Skip to content
llm-speed
← raw run
wrapped·Oct 17, 2026
/r/r_pob-co51bd-
your run

Qwen3.6-27B

on RTX 5090
75.1
tok/sdecode
RTX 5090 (32GB) + AMD EPYC 7543 32-Core Processor (120c) + 945GB
rank in tier
61/183RTX 5090 runstop 33%
best workload
chat-short
where the rig flew
slowest workload
agent-trace74.1 tok/s
where the rig struggled
backend
llama.cppllama.cpp is the universal backend — broad model support, modest speed ceiling.
faster than
  • phi-4-4bit on M3 Ultra74.4 tok/s
  • Qwen2.5-14B-Instruct-4bit on M3 Ultra70.9 tok/s
  • Qwen2.5-Coder-14B-Instruct-4bit on M3 Ultra70.5 tok/s
share
75.1 tok/s on RTX 5090 running Qwen3.6-27B — top 33% of RTX 5090 runs. Signed benchmark via llm-speed.
https://llm-speed.com/wrapped/r_pob-co51bd-
X / TwitterBlueskyMastodonRedditraw run

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.