Assessment Tiers
PTL assessment tiers are shorthand for where a composite PTL Score lands. The score is primary. The tier is the communication label.
Tier table
Section titled “Tier table”| Tier | Requirements | Composite |
|---|---|---|
| FRONTIER | All 5 scored engines | ≥ 0.85 |
| OPTIMIZED | 4 or more engines | ≥ 0.70 |
| CAPABLE | 3 or more engines | ≥ 0.60 |
| DEVELOPING | 2 or more engines including ACE | ≥ 0.45 |
| BASELINE | ACE only — first scored assessment | — |
| PENDING | No engines complete | — |
Scored engines are ACE, COOL, FLUX, PACE, and CORE. PROFILE is not scored. GRADE and ATLAS are not scored — they are the aggregation and recommendation layers.
Why tiers require multiple engines
Section titled “Why tiers require multiple engines”A single-engine score does not characterize an organization’s infrastructure comprehensively. ACE alone measures GPU utilization. An organization with excellent GPU utilization but a PUE of 1.8 and no carbon accounting is not FRONTIER. The tier system rewards organizations that allow measurement across multiple dimensions.
BASELINE is the exception: it recognizes that the first assessment often begins with sacct data alone, before facility data is available. BASELINE is a meaningful starting point, not a penalty.
Partial assessments
Section titled “Partial assessments”Organizations that have not yet provided data for all engines receive a score for the engines they have. A cluster assessed on ACE, PACE, and GRADE with a composite of 0.68 is CAPABLE, not penalized for missing COOL and FLUX. The assessment report discloses which engines were included.
This design rewards organizations that start the process rather than waiting until they can provide complete data.
Tier in assessment records
Section titled “Tier in assessment records”The tier appears in the report ID: PTL-YYYYMMDD-ORGSLUG-TIER. The report ID is tamper-evident — paired with a SHA256 hash of assessment fields. PTL can review any report ID on request.
The score is the truth
Section titled “The score is the truth”Two FRONTIER assessments — one scoring 0.851, one scoring 0.973 — are both FRONTIER. They are not the same. Readers who want precision should ask for the PTL Score, not only the tier. PTL always reports both.
Longitudinal record
Section titled “Longitudinal record”An organization that improves from BASELINE (0.42) to DEVELOPING (0.58) to CAPABLE (0.67) across three annual assessments has a record that tells a story. The tier sequence communicates the direction. The score sequence communicates the magnitude of improvement. Both matter.