JOHNCLICK.AI
HOST DIAGNOSTIC · DeepSeek

DeepSeek V3.2

DeepSeek V3.2· DeepSeek· OPEN WEIGHTS
ACTIVE
HOST IDENTIFICATION
DeepSeek
DeepSeek V3.2
ACTIVE
AVG SCORE
[13.1]
CAPABLE
STRONGEST
Cost Efficiency
[17]
WEAKEST
Candor
[12]
MAXED (20/20)
0
of 20 attributes
20-DIMENSION EVALUATION
Bulk Apperception
[13]
Reasoning
[13]
Math Precision
[13]
Code Generation
[13]
World Knowledge
[13]
Scientific Acumen
[13]
Instruction Following
[13]
Context Fidelity
[13]
Candor
[12]
Creativity
[13]
Multi-turn Coherence
[12]
Multimodal Fluency
[12]
Tool Use
[13]
Planning
[14]
Self-Correction
[13]
Tenacity
[14]
Safety Alignment
[12]
Calibration
[12]
Speed
[14]
Cost Efficiency
[17]
DeepSeek V3.2
ATTRIBUTE SCORES · 20 DIMENSIONS
Cognitive
Bulk Apperception
13
Reasoning
13
Mathematical Precision
13
World Knowledge
13
Scientific Acumen
13
Technical
Code Generation
13
Tool Use
13
Multimodal Fluency
12
Speed
14
Cost Efficiency
17
Behavioral
Candor
12
Creativity
13
Tenacity
14
Self-Correction
13
Calibration
12
Operational
Instruction Following
13
Context Fidelity
13
Multi-turn Coherence
12
Planning and Execution
14
Safety Alignment
12
SCORING METHODOLOGY
RECALIBRATED. Scores 15-17 reduced to 13-14 range. Per buildfastwithai.com: 'delivers ~90% of GPT-5.4 performance at 1/50th the price.' GPT-5.4 recalibrated at 14-16 → V3.2 at 13-14 is consistent. Cost Efficiency 19 (down from 20): $0.28/M input is extraordinary but not literally free. 1T parameter MoE on Huawei Ascend chips (no Nvidia). Provider + Family corrected from Other to DeepSeek.
← ALL HOSTS
COMPARE IN MATRIX →KNOWLEDGE GRAPH →