TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

benchmarks evals

The Company Benchmark Grew to 230,000 Documents

CorporateBench builds temporally consistent synthetic firms so question-answering systems can be tested at enterprise communication scale.

Published Updated Story ID: mp-2026-08-28-006
Read the complete editionStory JSON

Summary

CorporateBench builds temporally consistent synthetic firms so question-answering systems can be tested at enterprise communication scale.

Four generated firms range from 12 to 10,000 employees, with corpora exceeding 230,000 documents sampled from evolving knowledge bases. Five evaluated models performed worse as the inputs approached realistic scale. The benchmark preserves cross-document logical consistency, but synthetic firms remain a proxy for private organizations and their messier records.

Why it matters

CorporateBench builds temporally consistent synthetic firms so question-answering systems can be tested at enterprise communication scale.

Limits and context

  • The benchmark preserves cross-document logical consistency, but synthetic firms remain a proxy for private organizations and their messier records.

Key claims

  1. CorporateBench builds temporally consistent synthetic firms so question-answering systems can be tested at enterprise communication scale.

    Qualification: The benchmark preserves cross-document logical consistency, but synthetic firms remain a proxy for private organizations and their messier records.

    Evidence: source-2026-08-28-006

Sources

  1. arXiv preprint 2608.27391arXiv · primary research

Corrections

No corrections have been recorded for this story.