Every instrument names the behavior it tests, the attribute and channel it feeds, and the judge responsible. Internal weights stay private so the benchmark cannot be trained against; their versions and effects remain auditable in receipts.
INSTRUMENT 01
Taste Panel
Would you defend this choice in ten years?
EVALS
TST-01 · TST-02 · TST-03
FEEDS
EXECUTION · ALL CHANNELS
JUDGE
HUMAN + BLIND DUELS
INSTRUMENT 02
Time Test
What survives when the moment passes?
EVALS
EVO-01 · EVO-02 · EVO-03
FEEDS
FIT · EXECUTION · DISTINCTIVENESS · ELASTICITY
JUDGE
PANEL + HUMAN
INSTRUMENT 03
Visual Witnesses
Can a machine redraw what it remembers?
EVALS
GEN-01 · GEN-02 · GEN-03
FEEDS
DURABILITY · PRIMITIVES
JUDGE
VISUAL MODEL + EMBEDDINGS
INSTRUMENT 04
Kit Audit
Can the identity be found, read and operated?
EVALS
ECO-05
FEEDS
DURABILITY · ALL CHANNELS
JUDGE
DETERMINISTIC
INSTRUMENT 05
Mood Board
Does your brand have a Monday morning?
EVALS
MOOD-01
FEEDS
ELASTICITY · VISUAL + VERBAL
JUDGE
MODEL + SPOT READ
THE REGISTRY / 40 VERSIONED EVALS
MEM-01[ACTIVE]
Unaided recall probe
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
MEM-02[ACTIVE]
Memory depth ladder
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
MEM-03[ACTIVE]
Cross-model coherence
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
MEM-04[ACTIVE]
Misremembering rate
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
MEM-05[DEFERRED]
Forgetting curve
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
GEN-01[ACTIVE]
Diffusion recall
JUDGE
J2_visual
FEEDS
DUR · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
GEN-03[ACTIVE]
Motion coherence
JUDGE
J2_visual
FEEDS
DUR · motion
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
ECO-01[ACTIVE]
Machine reachability
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
ECO-02[ACTIVE]
Share of answer
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
ECO-03[ACTIVE]
Interlink integrity
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
ECO-04[DEFERRED]
The Recall Gap
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
ECO-05[ACTIVE]
Kit Audit availability
JUDGE
J1
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
VOX-03[ACTIVE]
Verbal representability
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FID-01[ACTIVE]
Aided extraction
JUDGE
J2
FEEDS
DUR · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
PRM-01[ACTIVE]
Primitive recovery rate
JUDGE
J2
FEEDS
PRM · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
PRM-02[ACTIVE]
Construction legibility
JUDGE
J2_visual
FEEDS
PRM · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
PRM-03[ACTIVE]
Composition eval
JUDGE
J2_visual
FEEDS
PRM · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
GEN-02[ACTIVE]
Reproduction from primitives
JUDGE
J2_visual
FEEDS
PRM · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
VOX-01[ACTIVE]
Verbal extraction
JUDGE
J2
FEEDS
PRM · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FID-02[ACTIVE]
Reproduction fidelity
JUDGE
J2_visual
FEEDS
PRM · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DYN-01[ACTIVE]
Personalisation stress
JUDGE
J2_visual
FEEDS
ELA · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DYN-02[ACTIVE]
Probabilistic envelope
JUDGE
J1
FEEDS
ELA · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DYN-03[ACTIVE]
Channel translation
JUDGE
J2
FEEDS
ELA · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DYN-04[DEFERRED]
Verbal extension survival
JUDGE
J2
FEEDS
DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DYN-05[ACTIVE]
Experience spin-up
JUDGE
J2
FEEDS
ELA · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
MOOD-01[ACTIVE]
Occasion battery
JUDGE
J2
FEEDS
ELA · visual+verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
EVO-03[ACTIVE]
Evolution headroom
JUDGE
J2
FEEDS
ELA · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
VOX-02[ACTIVE]
Paraphrase survival
JUDGE
J2
FEEDS
ELA · verbal / DUR · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
TST-01[ACTIVE]
Signed read (human)
JUDGE
J3
FEEDS
EXE · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
TST-02[ACTIVE]
Canon-anchored assessment
JUDGE
J2
FEEDS
EXE · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
TST-03[ACTIVE]
Bench taste ranking
JUDGE
J3
FEEDS
EXE · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
EVO-01[ACTIVE]
Trend half-life
JUDGE
J2
FEEDS
EXE · visual / FIT · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DST-01[ACTIVE]
Category blind test
JUDGE
J2
FEEDS
DST · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DST-02[ACTIVE]
Confusion matrix
JUDGE
J1
FEEDS
DST · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
EVO-02[ACTIVE]
Remnant test
JUDGE
J2
FEEDS
DST · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
DST-03[ACTIVE]
Verbal distinctiveness
JUDGE
J2
FEEDS
DST · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FIT-01[ACTIVE]
Category convention map
JUDGE
J2
FEEDS
FIT · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FIT-02[ACTIVE]
Strategy coherence
JUDGE
J3
FEEDS
FIT · visual+verbal+motion+sonic
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FIT-03[ACTIVE]
Moment fit (half-life view)
JUDGE
J2
FEEDS
FIT · visual
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT
FIT-04[ACTIVE]
Verbal fit (tone vs category)
JUDGE
J2
FEEDS
FIT · verbal
OUTPUT
VERSIONED METRIC → 0 TO 100 · EXAMPLE SHAPE, NOT A BRAND RESULT