JOHNCLICK.AI
HOST DIAGNOSTIC · Meta

Llama 4 Scout

Llama 4 Scout · ACTIVE· Llama· CLOSED
ACTIVE
ATTRIBUTE MATRIX · ATRBT GROUP 01
Bulk Apperception
[12]
Reasoning
[12]
Math Precision
[11]
Code Generation
[12]
World Knowledge
[12]
Scientific Acumen
[11]
Instruction Following
[12]
Context Fidelity
[16]
Candor
[11]
Creativity
[11]
Multi-turn Coherence
[11]
Multimodal Fluency
[12]
Tool Use
[11]
Planning
[11]
Self-Correction
[10]
Tenacity
[10]
Safety Alignment
[11]
Calibration
[11]
Speed
[16]
Cost Efficiency
[18]
HOST TELEMETRY
AVG
12.1
/20
BEST
18
Cost Efficiency
WEAK
10
Self-Correction
PARAMETERS
Undisclosed
CONTEXT
—
PRICING
—
STATUS
ACTIVE
ATTRIBUTE SCORES · 20 DIMENSIONS
Cognitive
Bulk Apperce…
12
Reasoning
12
Mathematical…
11
World Knowle…
12
Scientific A…
11
Technical
Code Generat…
12
Tool Use
11
Multimodal F…
12
Speed
16
Cost Efficie…
18
Behavioral
Candor
11
Creativity
11
Tenacity
10
Self-Correct…
10
Calibration
11
Operational
Instruction …
12
Context Fide…
16
Multi-turn C…
11
Planning and…
11
Safety Align…
11
SCORING METHODOLOGY
RECALIBRATED per ADR-NM108 (April 17, 2026). Gen 1 scored Scout HIGHER than Maverick (15-16 vs 7-8) which is backwards — Scout is the smaller, efficiency-focused variant. Recalibrated to 11-12 range. Context Fidelity kept high at 16 (10M token window is genuinely exceptional). Cost 18 + Speed 16 validated (efficiency is Scout's value prop). Open weights, part of Llama 4 family.
← ALL HOSTS
COMPARE IN MATRIX →KNOWLEDGE GRAPH →