Alpha Evals

Independent judgment on AI governance.

Alpha separates governance quality from governance materiality, keeps every opinion tied to a declared evidence basis, and publishes limitations alongside the judgment.

Open sample ratingRequest Evals access

Representative instrument

Eval record

No issuer selected
Verdict
NRNot Rated is an explicit coverage state, not a low score.
Evidence basis
UnassignedPublic Information or Verified Assessment must be declared.
Method
The Alpha Standard Version 1Public calibration candidate
Governance quality
Alpha AIGR
Materiality
Alpha GMI
Scale
AAA / D / NR
Current standard state
Public calibration candidate

The Eval system

Verdict, evidence, methodology, consequence.

A rating should make the opinion inspectable without collapsing quality, materiality, source coverage, and analytical limits into one opaque score.

1Alpha AIGRThe independent opinion on how effectively an organization governs AI.Quality2Alpha GMIThe companion view of how material AI governance is to the organization.Materiality3The Alpha StandardSix weighted pillars, controlled evidence bases, public score anchors, and one rating scale.Method4IndependenceMembership, access, education, and participation do not influence a rating opinion.Control

Rating scale

The published rating scale.

Letter grades express Alpha's rating opinion. NR is a distinct coverage state and never represents a low score.

Rating grades
AAA / AA / A / BBB / BB / B / CCC / CC / C / DAvailable
Coverage state
NR · Not RatedLimited availability

Evidence basis

A single scale with governed evidence boundaries.

Version 1 is published for review and calibration and is not yet the active production rating engine. Historical opinions retain their original methodology version.

Alpha AIGR, Public Information
Based only on attributable public evidence available by the evidence cutoff date. Normally capped at A, or AA with disclosed independent assurance. Never AAA.Limited availability
Alpha AIGR, Verified Assessment
Based on evidence verified by Alpha within a signed assessment perimeter. Eligible for the full AAA to D scale when every publication gate is met.Limited availability

Coverage architecture

Comparable records without collapsing unlike subjects.

Organizations and vendors may receive governed rating opinions. Models, agents, workflows, and people retain subject-appropriate evaluation scales and explicit authority boundaries.

1EnterprisesGovernance quality, material exposure, control effectiveness, evidence confidence, and rating history.AAA–D / NR
2Frontier labs and vendorsGovernance posture, dependencies, incidents, assurance, and institutional consequence.AAA–D / NR
3Model configurationsReproducible evaluations of exact provider, version, product surface, tools, prompts, and harness settings.Benchmark score4People and workflowsReadiness, reliability, escalation, and authority integrity remain separate from issuer-style ratings.Assessment

Decision use

Independent signals for consequential oversight.

BoardsOversight posture, committee reporting, and governance-gap prioritization.
InvestorsPeer comparison, stewardship focus, and diligence on material AI exposure.
InsurersUnderwriting context for governance controls, incident readiness, and accountability.
Public institutionsComparable reference points for supervision, policy analysis, and market monitoring.