Vibeleaderboard
← All Intel
Intel / post

Hermes has twelve browser tools. Browser Use mode replaces them with a single on

Source
x.com
Date
Nous Research@NousResearch

Hermes has twelve browser tools. Browser Use mode replaces them with a single one, driven by @browser_use's CLI 3.0. Instead of a dozen schemas in every request and a tool call per click, the agent writes a script. In our tests that cut token use 48-66% with no accuracy drop.

Context

Nous Research changed how its Hermes operates a browser: Browser Use mode replaces twelve separate browser tools with one, driven by version 3.0 of Browser Use's command-line tool, through which the agent writes a script. This removes a dozen tool descriptions from every request and lets several actions be expressed together instead of requiring a separate call for each click. The announcement reports 48–66% lower use without an accuracy drop in its tests.

The benchmark documentation gives a narrower picture: savings were largest on harder multi-step tasks (about 60–66% for two models), around 30% for one model on easy tasks, and roughly none for a smaller coding model. Tests used a small set of scraping-practice websites with stable content and no anti-bot defenses, with three repeats per task, and the aggregate results were recovered from session transcripts after the raw per-run logs were lost. The supported conclusion is that changing the tool interface can reduce the text an agent processes, with task-dependent savings.

Terms in this piece · Glossary
  • AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
  • token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
  • benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
More from Nous Research
Recommended reads
Comments

Checking sign-in…

Loading comments…