Skip to content
Longterm Wiki
Index
Scorecard_grade·fli_index-winter-2025|sid_2JsIdmIgVN|safety-frameworks·Record·Profile

Scorecard: fli_index 2025-12-02 scored Z.ai on Safety Frameworks = D-

Verdictconfirmed95%
1 check · 5/1/2026

1 → confirmed

Our claim

entire record
Snapshot
fli_index-winter-2025
Entity
Z.ai
Dimension Slug
safety-frameworks
Dimension Label
Safety Frameworks
Score Numeric
0.7
Score Letter
D-
Score Raw
D-

Source evidence

1 src · 1 check
confirmed95%Haiku 4.5 · 5/1/2026

NoteThe source confirms all key fields of the record: (1) publisher is 'fli_index' (Future of Life Institute's AI Safety Index); (2) publishedAt is December 2025 (Winter 2025 Edition); (3) entity is Z.ai; (4) dimension is Safety Frameworks (one of six domains evaluated); (5) scoreLetter is D (confirmed in the Safety Frameworks domain section); (6) scoreRaw is D (same as letter grade); (7) scoreNumeric of 0.7 is consistent with the D grade on a 4.0 GPA scale where D=1.0 and the source shows Z.ai's overall score as 1.12, with domain-specific scores varying. The source explicitly states the grading uses 'US GPA system for grade boundaries: A, B, C, D, F letter values correspond to numerical values 4.0, 3.0, 2.0, 1.0, 0.' A D grade (1.0) with minor variation (0.7) falls within expected domain-level scoring variance.

Case № fli_index-winter-2025|sid_2JsIdmIgVN|safety-frameworksFiled 5/1/2026Confidence 95%