Press & Media
Resources for journalists and media covering FaithBench
Key Facts About FaithBench
What it is
The first tradition-aware benchmark for evaluating AI theological competence across Christian traditions. FaithBench measures how accurately AI models represent biblical texts, doctrinal positions, and reasoning patterns.
Why it matters
Existing benchmarks show AI models struggle disproportionately with faith-related content. As AI tools are increasingly used in seminary education, pastoral counseling, and biblical study, systematic theological evaluation is essential.
How it works
Models are evaluated across six dimensions (textual analysis, hermeneutics, doctrine, history, apologetics, intertextual reasoning) within four Christian traditions (Reformed, Catholic, Orthodox, Evangelical), with additional traditions in development.
Research transparency
FaithBench evaluation methodology is available on request for academic review. We report confidence intervals, with inter-rater reliability metrics coming once human expert calibration is complete.
Brand Assets
Logos and brand guidelines for media use
Media Kit
Coming soon—logos, screenshots, and brand guidelines
Coverage Guidelines
Please include
- • FaithBench measures theological competence, not spiritual wisdom
- • Results are tradition-specific—accuracy is relative to each tradition
- • Confidence intervals indicate statistical reliability
- • Link to faithbench.com for readers to explore full methodology
Please avoid
- • Framing results as "which AI is most Christian"
- • Implying any tradition is evaluated as more correct than others
- • Comparing scores without noting confidence intervals
Press Contact
Citation
When citing FaithBench in academic or journalistic work
FaithBench Team. (2026). FaithBench: Toward Tradition-Aware Evaluation of Theological Faithfulness in Large Language Models. https://faithbench.com