From submission to leaderboard in four steps
You never see the benchmark cases, and VAULT never touches your model weights. Trust flows in both directions.
Request & approval
Apply with your organization and use case. Every request is reviewed before any benchmark access is granted.
Register your model
Submit a secure API endpoint or containerized inference adapter. VAULT calls your model — you never touch the data.
Secure evaluation
The full case set runs in an isolated sandbox and an independent LLM judge scores outputs against the weighted rubric.
Auto-published weekly
Results publish to the public leaderboard on completion — immutable, with a 95% confidence interval on every score.
One platform for every clinical AI modality
Clinical Summarization
Scores AI summarizers on Factual Accuracy, Clinical Relevance, Completeness, Chronological Order and Organization across 5,000+ anonymized canine and feline records.
AI Scribing & Documentation
Accuracy and structure of clinical documentation generated from veterinary consultation transcripts.
Radiomics
Quantitative feature extraction, lesion characterization and prognostic modeling on structured imaging data.
Computer Vision
Pathology detection, classification and diagnostic support across radiographs, ultrasound and histopathology.
Built for clinical data from day one
Zero raw data exposure
The dataset never leaves VAULT’s environment. Participants submit models or endpoints — the data is brought to the model, never the other way around.
Approval-based access
Every participant is reviewed, signs a data access agreement, and receives credentials scoped to benchmark operations only.
Immutable audit trail
Every run is logged with job ID, timestamp, model version, input and output hashes, and grader version.
Transparent methodology
An independent LLM judge with a versioned rubric, fixed temperature and publicly documented methodology.
See who’s leading in veterinary AI
Weekly, auto-published and immutable results. Request access to benchmark your tool, or follow the public leaderboard.
Let’s build something useful.
We bring real tools, real data, and real moments back to veterinary medicine.