Omni EnterpriseRAG-Bench: Agentic RAG Evaluation Results and Artifacts

Omni's public verification artifacts for the full-500 agentic benchmark run of EnterpriseRAG-Bench. The dataset includes answer files, evaluation result files, patch manifests, and a merge script to verify reported metrics against benchmark questions and gold metadata. It was published by the author 'getomnico' and last updated on May 21, 2026.

Use Cases

• Verify reported performance metrics for RAG systems based on the provided answer and evaluation files.
• Reproduce benchmark results using the included merge script and patch manifests.
• Analyze agentic RAG performance on enterprise-oriented questions based on the benchmark corpus.
• Compare different RAG system outputs against a standardized evaluation framework.

Strengths

• Includes artifacts for a full-500 agentic benchmark run, suggesting a substantial evaluation scale.
• Provides a merge script for independent verification of reported metrics.
• Contains patch manifests, which may indicate version control and reproducibility efforts.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset does not re-host the full benchmark corpus, limiting standalone analysis.

Provenance

Source

getomnico (Omni) via Hugging Face Datasets.

Collection Method

Likely generated as part of running the EnterpriseRAG-Bench evaluation.

Freshness

Last updated 2026-05-21 05:14:02; freshness should be verified.