Observed arrival · 2026-10-24

LocalAgent Board’s Hermes Agent leaderboard

localagentboard.com Visit website

Editorial interest 78/100 Selection signal · not a rating of the site

A hands-on leaderboard compares local language-model setups on Hermes Agent tasks, speed, and memory use.

Landing page captured for the 2026-10-24 issue.

For: People running Hermes Agent on local hardware
Worth noticing: Only 16 of 22 setups are tested; nine near-leaders within two points are treated as ties because reruns vary by about ±2.

Field notes

Each setup is identified by model, quantization, engine, memory use, speed, and task score, with detail links and copy-config controls. The stated score comes from 10 real agent tasks attempted three times apiece; tool-call success is reported separately, alongside the full-run date and Hermes version. Testing is underway rather than complete: 16 of 22 setups are listed, and the page warns that repeat runs vary by about two points.

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

○OpenPublic substance visible

✦PrettyNotable craft visible

●ProPolished or operationally mature

◎NicheUnusually specific use

One card from the complete issue

Pageview Provinces

506,715 arrived 1,000 judged 975 catalogued Enter the complete issue