Vendor-neutral benchmark suite for multi-model databases with AI/LLM and RAG evaluation.
372
CHIMERA Suite is a vendor-neutral, scientifically rigorous benchmark framework for modern multi-model databases and AI-enabled data systems.
It evaluates relational, graph, vector, and document workloads, including hybrid scenarios, and adds built-in support for LLM and RAG quality assessment. The project focuses on fair comparison, transparent methodology, and reproducible results.
Key capabilities:
Multi-model benchmark coverage: relational, graph, vector, document, hybrid AI/LLM and RAG evaluation: faithfulness, relevance, hallucination checks, citation quality Statistical rigor: confidence intervals, significance testing, effect sizes, outlier handling Vendor-neutral reporting: no branding bias, neutral ordering, accessible visual design Multiple output formats for analysis and reporting workflows Why CHIMERA:
Fair cross-system comparison without vendor favoritism Standards-oriented benchmark reporting Ready for CI/CD benchmark pipelines and repeatable regression tracking Typical use cases:
Compare database systems under identical workloads Track performance regressions over time Validate RAG/LLM output quality with objective metrics Produce transparent benchmark reports for engineering and research teams License: MIT License
Content type
Image
Digest
sha256:e96ddbd71…
Size
223.8 MB
Last updated
7 months ago
docker pull themisdb/chimera