Our models
Low-latency, high precision claim extraction and grounding models with deterministic post-verification to ensure traceable, high-reliability AI outputs.
Flash 1
Our fastest, most affordable model for real-time, lower stakes use cases.
Latency: < 1 seconds per claim
Web grounding
Ground to PDF, office documents, images, audio, video + more
Claim theme extraction for lowest latency
Ideal for AI co-pilots and assistants, content review, sales enablement automation and internal QA
Pro 1
Our balanced model with low latency and high claim granularity for most applications
Latency: < 5 seconds per claim
Web grounding
Ground to PDF, office documents, images, audio, video + more
Deterministic verification
Atomic claim extraction for balanced accuracy and latency
Full traceability from output to source for auditability and regulatory compliance
Ideal for regulated deployments in finance, legal and healthcare, enterprise report generation and validation, contract review and compliance checks, model oversight and governance
Ultra 1
Our highest claim granularity with expansive grounding effort for critical, zero-miss, high-stakes uses.
Latency: < 30 seconds per claim
Web grounding
Ground to PDF, office documents, images, audio, video + more
Deterministic verification
Deepest atomic claim extraction for highest precision
Zero-error tolerance with in-built self-checks
Full traceability from output to source for auditability and regulatory compliance
Ideal for clinical decision support systems, legal advice and filings, board or public company communications, indemnified and regulated workflows