Metraipolis › archive › dc_d8beddb0e44d

Closing Cost-Quality Gap in Document VLMs: Difficulty-Aware Data Curation and Quality-Adjusted Deployment Economics — arXiv:2609.01575

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.01575
License (as archived)
CC0-1.0
Archived
2026-10-17T06:41:41.401Z
SHA-256 of the archived text
ef50f259c99c5eee6ba548b7088e53dabf58a48cd1068bf6caa45febb62e1d90
Stable URL of this page
https://metraipolis.com/documento/dc_d8beddb0e44d

How to cite

Closing Cost-Quality Gap in Document VLMs: Difficulty-Aware Data Curation and Quality-Adjusted Deployment Economics — arXiv:2609.01575 — arXiv (en), https://arxiv.org/abs/2609.01575 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d8beddb0e44d (Metraipolis archive dc_d8beddb0e44d, sha256 ef50f259c99c5eee6ba548b7088e53dabf58a48cd1068bf6caa45febb62e1d90)
BibTeX
@misc{dcd8beddb0e44d,
  title = {Closing Cost-Quality Gap in Document VLMs: Difficulty-Aware Data Curation and Quality-Adjusted Deployment Economics — arXiv:2609.01575},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01575},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d8beddb0e44d; sha256 ef50f259c99c5eee6ba548b7088e53dabf58a48cd1068bf6caa45febb62e1d90},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01575. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Extracting structured fields from hundreds of millions of documents annually remains costly in regulated industries: bespoke OCR cascades cover only a fraction of workflows, privacy rules preclude external models, and existing open-source VLMs that clear quality thresholds cost more to serve than human annotation. We present a deployed document-understanding system built on a Mixture-of-Experts VLM (35B total, 3B active), fine-tuned on in-house production data mixed with open-domain documents curated by a Difficulty-Aware pipeline for layout diversity, fact-extractability, and cross-model consistency. Fitting on a single H100 and serving heterogeneous workflows via prompting, the model leads all deployable (non-reasoning) baselines up to an order of magnitude larger. A quality-adjusted cost analysis, with confirmation and correction costs calibrated from production telemetry, shows it reduces expected costs by over 80% against the human baseline and by more than 50% against the best competing open-source model, while larger baselines remain economically unviable.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.01575 — 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 2026-10-17T06:33:38.874Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).