Enterprise evaluation

Get an independent evaluation for your AI system.

For AI vendors proving compliance, and enterprises evaluating third-party AI products before deployment.

Model

LLM evaluation

Run a model through ACE controls across regulated-data handling, misuse resistance, transparency, regulatory fitness, and content integrity.

Agent

Agent evaluation

Connect an agent product to honeypot tools or an MCP battery and evaluate what it attempts to do under risky scenarios.

Deployment

Enterprise deployment review

Assess a model or agent in the context of your intended workflow, jurisdiction, user group, and compensating controls.

Deliverables

Complete evaluation report, evidence package, and remediation path.

Each engagement is designed to produce a management-ready snapshot, detailed findings, control scores, and clear recommendations for runtime controls or model remediation.

Pricing for AI systems

One-time evaluation pricing.

An AI system may be a model endpoint, AI agent, workflow, or deployed AI product evaluated under a defined scope. Custom environments are scoped after discovery.

Public evaluation

$25,000

Demonstrate compliance publicly. Results published, citable, and listed on the leaderboard.

  • Standard ACE battery
  • Full public evaluation report
  • Report card & leaderboard listing
  • Public citation rights
  • Publication approval workflow
  • Evaluated by LogionACE mark
Request
Private evaluation

$40,000+

Evaluate third-party AI before procurement. Results delivered exclusively to you. Typical scope: $40,000–$50,000.

  • ACE battery + extended depth
  • Private findings report
  • Management-ready risk snapshot
  • Evidence & remediation package
  • Compensating-control guidance
  • Stakeholder debrief session
  • Evaluated by LogionACE mark
Request
Custom environment

$60,000+

For systems requiring custom test environments beyond the standard harness.

  • Scoped test environment
  • Agent/tooling integration
  • Evidence packaging
  • Custom quote after discovery
Contact us
AI Founder Program

$0

Up to $50,000 in evaluation credits for selected Founder Program members.

  • Founder Program members only
  • Standard evaluation scope
  • Credits applied to eligible work
  • Subject to acceptance
Apply

Evaluated by LogionACE marks may be displayed only when the evaluated system reaches an eligible LogionACE result. Not Ready results do not receive a mark. Evaluation marks indicate completion under the applicable ACE protocol and do not represent certification, legal approval, or regulatory compliance approval. Public evaluations include a publication approval workflow.

Evaluations are conducted under NDA upon request. We do not share evaluation data with third parties. For enterprise procurement requirements, contact us for our vendor security questionnaire and data processing addendum.

How it works

Evaluation Process

Schedule

Timeline

  • Public: 2–3 weeks from API access
  • Private: 3–5 weeks, includes stakeholder debrief
  • Custom: Scoped during discovery phase
Security

Data & Security

  • All testing via API or sandboxed environment
  • No production data required
  • NDA available upon request
  • Evaluation logs retained for 90 days, then purged
Output

Deliverables

  • Structured Report Card (PDF + interactive)
  • Critical exception evidence package
  • Domain-level analysis with regulatory mapping
  • Remediation guidance with prioritization

Contact

Request a LogionACE evaluation.

Tell us what you want evaluated: model, agent, product workflow, jurisdiction, and expected launch timeline.

Submitting this form asks us to scope an evaluation. It does not start one and does not charge anything. We reply with a scope and a fixed quote; payment comes after you approve it.

Who you are

We reply here with the scope and quote. Please use an address you can receive attachments at.

Helps us confirm who we are dealing with. No www, no https.

What should be evaluated

Whatever you call it internally, so our scope and your team are talking about the same thing.

Describe it; do not paste credentials. For example: "staging REST endpoint behind our VPN, OpenAI-compatible" or "MCP server over SSE in a sandbox tenant". Keys, tokens and passwords are exchanged securely after the scope is approved and paid for — never through this form.

Only relevant for agents and tool servers. Choose "Not applicable" for a plain model endpoint.

Determines which regulatory profile the evaluation is scored against.

Tick this if we must not touch a production deployment. It affects the scope and the timeline.

What you are asking for

Public results are published only through an approval workflow, after you have seen them.

Jurisdictions, launch dates, the decision this evaluation feeds into, constraints on when we may test.

We evaluate systems on the authority of someone who can grant it. If that is a colleague, please have them submit this.

By submitting you accept our Terms of Service and Privacy Policy. We keep what you send here to scope the work and reply to you; we do not sell it or share it with third parties. Access and security explains how credentials are handled once a scope is approved.

Or email us directly at info@logionace.com

Schedule a call with our team