# Harness-Zero raises macro task success to 44.3% without the specialized harness

A paper from Google and colleagues describes Harness-Zero, which uses an optimized agent harness only during training. After that harness is removed, macro task success rises from 23.3% to 44.3%, above the 41.7% reached by the base model with the harness attached.

Language: en
Time zone: America/Los_Angeles
Current time (UTC, cached as of): 2026-11-10T06:09:00.000Z
Query date (America/Los_Angeles): 2026-11-09

HTML: https://didcodexreset.com/news/3762ba79953a4209984621f6.html

Content language: en
Localization state: sameLanguage

Source: elvis · Published 9/24/2026, 05:30:34

A paper from Google and colleagues studies whether an agent harness can be distilled into a model. Harness-Zero keeps the optimized harness in training only. Because the optimized harness and the deployment harness use different action spaces, a harnessing agent guided by the optimized harness corrects the student's responses in the deployment action space before they run, and those corrected runs become the training demonstrations.

With the specialized harness removed, macro task success goes from 23.3% to 44.3%, higher than the 41.7% the base model reaches with the harness attached. Across 28 harness-induced behaviors in knowledge work, tool use, and science, 82.3% are recovered on average. For frontier models using the same evolved harness, the agent-as-harness form also beats the code-as-harness form. The paper is [arXiv:2609.24974](https://arxiv.org/abs/2609.24974).

Tags: Harness-Zero, Agent harness, Google, AI research

[View original post](https://x.com/omarsar0/status/2103095360239636666)