← Back to DizyDiz

LLM Benchmark Archive

Head-to-head benchmarks of the models running our homelab agents (Mr. Peepers, Velma, Louis). Same prompts, same parameters, real hardware. Updated automatically when new benches run.
loading… · bench runs · swap decisions

Speed over time tok/s by model & host

Internal IPs, credentials, and full thinking traces are stripped from sample outputs. Original NDJSON archive lives on Hommer for ops use.

Decisions

All runs

Date Model Host Role tok/s (med) Wall (s) Sample outputs