Supported Foundation Models

This table shows the foundation models that SLI's Legal Intelligence Assistants (LIAs) can work with as of each "Last Updated" date. These are also the only AI platforms we use internally for service delivery through our proprietary LIAs. If you're a customer with an active commercial subscription to any of these providers, your SLI-developed LIAs can leverage that existing relationship.

We maintain this list to provide transparency about our technology ecosystem, but the AI landscape changes rapidly. Data retention policies, security certifications, and deployment options may have evolved since our last verification.

We make reasonable efforts to keep this information current, but you should always consult the model providers' official documentation for the most up-to-date details.

This information is provided purely for reference and does not constitute legal, compliance, or technical advice. Every organization has unique requirements that might affect which foundation models are appropriate for your specific use case. You should consult your legal counsel. IT, and security teams when making decisions about LLM implementations.

In line with our Privacy Statement, we apply careful safeguards when using these or any AI tools. If you have questions about how we work with these foundation models or our data handling practices, please see our FAQ for more details or contact us at info@sl-intelligence.com.

Provider Data Retention Policy Use of Data for Training (Paid Business Accounts) Deployment Method Security & Compliance Highlights Last Updated
OpenAI 30 days default retention Data not used for training by default Cloud only SOC 2 Type II compliant April 4, 2025
Anthropic 30 days backend deletion Data not used for training by default Virtual Private Cloud SOC 2 Type I & II compliant
ISO 42001
HIPAA capable
April 4, 2025
Google No long-term storage, 72 hours backend Data not used for training by default Fully managed by Cloud with VPC ISO 27001
SOC 1-3 compliant
HIPAA
FedRAMP
April 4, 2025