Metraipolis › archive › dc_56a6eaaa0250
Compact-Bench: Measuring What Context Compaction Costs Long-Horizon Coding Agents — arXiv:2610.31710
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.31710
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-26T19:50:54.531Z
- SHA-256 of the archived text
aea215e41061425f74b9017e051737487b96c27c683cc4836e4805ea0716a504- Stable URL of this page
https://metraipolis.com/documento/dc_56a6eaaa0250
How to cite
Compact-Bench: Measuring What Context Compaction Costs Long-Horizon Coding Agents — arXiv:2610.31710 — arXiv (en), https://arxiv.org/abs/2610.31710 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_56a6eaaa0250 (Metraipolis archive dc_56a6eaaa0250, sha256 aea215e41061425f74b9017e051737487b96c27c683cc4836e4805ea0716a504)
BibTeX
@misc{dc56a6eaaa0250,
title = {Compact-Bench: Measuring What Context Compaction Costs Long-Horizon Coding Agents — arXiv:2610.31710},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.31710},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_56a6eaaa0250; sha256 aea215e41061425f74b9017e051737487b96c27c683cc4836e4805ea0716a504},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31710. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Context compaction lets long-horizon coding agents continue beyond their context windows, but can erase the state needed to finish a task. Prior evaluations measure efficiency, retained information, or end-to-end success, leaving the effect of a specific history replacement difficult to isolate. Compact-Bench pairs full-history and summary-only continuations from a shared repository state, and evaluates recurring details and successive production compactions. It covers 35 SWE-bench repairs, 146 agentic sessions, and 106 production chains. Across 14 summarizer backbones, stable damage ranges from 5-64% of 22 full-history successes: Claude Opus 4.8 preserves the most repairs, while DeepSeek-Flash damages the most. With Kimi-K3 as summarizer, the Kimi Code prompt preserves the most task continuations (19/22), whereas Claude Code retains the most recurring details (15.0%). LLM summaries retain at most 16.0% of recurring details, compared with 79.1% for mechanical extraction; production summaries retain 15.7% at the first boundary. Failures recur across methods: one repair is damaged by 13/14 backbones, and another by all 5 prompts. These results show that reliable compaction requires preserving the progress needed to complete a task, and that retaining relevant information alone does not ensure successful continuation. Code is available at this https URL and data is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Software Engineering (cs.SE) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.31710 — 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 2026-10-26T19:45:33.350Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).