Metraipolis › archive › dc_d5ff0af151db

ParetoSaddler: Pareto-Guided Optimization of Agent Harnesses for Quality and Cost — arXiv:2611.02079

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.02079
License (as archived): CC0-1.0
Archived: 2026-11-03T10:59:31.821Z
SHA-256 of the archived text: fb1a5f195df595ec033b88660dcd47be73a6a5fc348e74c5df3fcd2d2227152d
Stable URL of this page: https://metraipolis.com/documento/dc_d5ff0af151db

How to cite

ParetoSaddler: Pareto-Guided Optimization of Agent Harnesses for Quality and Cost — arXiv:2611.02079 — arXiv (en), https://arxiv.org/abs/2611.02079 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d5ff0af151db (Metraipolis archive dc_d5ff0af151db, sha256 fb1a5f195df595ec033b88660dcd47be73a6a5fc348e74c5df3fcd2d2227152d)

BibTeX

@misc{dcd5ff0af151db,
  title = {ParetoSaddler: Pareto-Guided Optimization of Agent Harnesses for Quality and Cost — arXiv:2611.02079},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02079},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d5ff0af151db; sha256 fb1a5f195df595ec033b88660dcd47be73a6a5fc348e74c5df3fcd2d2227152d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02079. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Harness optimization can substantially improve LLM agents without modifying model weights, but existing methods primarily optimize response quality. Extending them to joint quality--cost optimization introduces a challenge: applying the deployment preference during intermediate updates can prematurely reject candidates that are valuable for later improvements. We introduce ParetoSaddler, which addresses this mismatch by separating intermediate candidate promotion from final deployment selection. At each iteration, an optimizer updates a parent harness to produce a candidate, and then compares the two over repeated executions and uses Bayesian bootstrap Pareto support to promote candidates showing credible progress in either quality or cost. Promoted candidates receive broader development evaluation and remain available as parents for later optimization steps. We evaluate ParetoSaddler on ClawMark and GAIA2 using domain-disjoint splits. On ClawMark, ParetoSaddler improves quality from 70.86\% to 81.35\% while reducing cost from \$1.17 to \$1.01 per task; on GAIA2, it improves quality from 56.22\% to 62.89\% while reducing cost from \$0.78 to \$0.45 per task. The Pareto-selected harnesses also achieve higher quality at lower cost than quality-only harness optimization. Analysis shows that Pareto-aware optimization retains intermediate candidates that would be discarded by deployment preference but can support later improvements, while Bayesian bootstrap screening filters out candidates whose gains are weakly supported across runs.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2611.02079 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-03T10:59:22.309Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).