Terms of Use
Effective Date: August 15, 2026By accessing TrustTheBench, you agree to use the information for lawful research, educational, and benchmarking purposes. All data is provided "as is" with transparent methodology, source attribution, and forensic integrity.
1. Purpose & Educational Use
TrustTheBench is an open-access platform intended to assist AI researchers, engineers, policymakers, and practitioners in understanding the scope, limitations, contamination risks, and saturation states of foundation model benchmarks.
You are free to view, reference, and cite benchmark specifications, trust scores, and model performance timelines in your academic papers, technical reports, and commercial evaluations, provided you credit TrustTheBench or the underlying primary authors.
2. Disclaimer of Warranties
The platform and all included benchmark metadata, model evaluations, and score calculations are provided on an "as is" and "as available" basis without warranties of any kind.
While we make every effort to ensure reproducibility, contamination detection, and verified scoring, model performance on real-world downstream tasks may diverge from academic benchmark metrics.
3. Contact Us
If you have any questions regarding these Terms of Use, please contact us at terms@trustthebench.com or visit our Contact Page.