Google paper shows distilling the agent harness lifts task success from 23.3% to 44.3%
omarsar0 · x · 2026-09-24
A new paper from Google and colleagues studies whether an agent harness can be distilled into the model itself.
- With the specialized harness removed, macro task success jumps from 23.3% to 44.3% — above the base model's 41.7% even with the harness attached.
- Harness-Zero uses the optimized harness only during training: since its action space differs from the deployment harness, a harnessing agent corrects the student's responses before they run, and those corrected runs become training demonstrations.
- 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 code-as-harness.
Robustness remains to be seen, but the approach is notably interesting.
More from coding & agent
- Why inference speed will be AI products' biggest competitive edge in 1–2 years — itsOmSarraf_ · 2026-11-11
- What happened when I moved vector retrieval into Snowflake SQL instead of orchestrating RAG outside — dima806_dima · 2026-11-11
- OpenAI engineer runs Codex for 26 days straight hunting Atlantis, spawning "vibe archaeology" — Polymarket · 2026-11-11
- He replaced a $5,000/month ghostwriter with 11 Claude Skills and earned $47,000 from one post — aryanXmahajan · 2026-11-11
- Refero Styles ships 2,000+ DESIGN.md files to give Claude Code real design taste — bbssppllvv · 2026-11-11
- Agent circuit breaker: openclaw shorts bot halts itself after 3 failed experiments — Rich_Standard_1724 · 2026-11-11