HOST DIAGNOSTIC · Google

Gemini 3.1 Pro

Gemini 3.1 Pro · ACTIVE· Gemini· CLOSED
ACTIVE
ATTRIBUTE MATRIX · ATRBT GROUP 01
Bulk Apperception
[18]
Reasoning
[16]
Math Precision
[17]
Code Generation
[16]
World Knowledge
[17]
Scientific Acumen
[19]
Instruction Following
[16]
Context Fidelity
[17]
Candor
[15]
Creativity
[15]
Multi-turn Coherence
[16]
Multimodal Fluency
[18]
Tool Use
[16]
Planning
[16]
Self-Correction
[15]
Tenacity
[15]
Safety Alignment
[15]
Calibration
[15]
Speed
[15]
Cost Efficiency
[15]
HOST TELEMETRY
AVG
16.1
/20
BEST
19
Scientific Acumen
WEAK
15
Candor
PARAMETERS
Undisclosed
CONTEXT
PRICING
STATUS
ACTIVE
ATTRIBUTE SCORES · 20 DIMENSIONS
Cognitive
Bulk Apperce…
18
Reasoning
16
Mathematical…
17
World Knowle…
17
Scientific A…
19
Technical
Code Generat…
16
Tool Use
16
Multimodal F…
18
Speed
15
Cost Efficie…
15
Behavioral
Candor
15
Creativity
15
Tenacity
15
Self-Correct…
15
Calibration
15
Operational
Instruction …
16
Context Fide…
17
Multi-turn C…
16
Planning and…
16
Safety Align…
15
SCORING METHODOLOGY
RECALIBRATED per ADR-NM108 (April 17, 2026). Prior Gen 1 scores were 17-20 across the board; Arena Text 1493 maps to 15-17 band, not 19. Adjusted down 2-3 pts on most dimensions. Scientific Acumen kept at 19 (GPQA Diamond 94.3% = #1 independent). Key benchmarks: AAII 57 (tied #1) → Bulk 18. Arena Text #2 at 1493 → Reasoning 16. SWE-Bench 78.80% (#1) → Code 16. ARC-AGI-2 77.1% (2x predecessor) → Planning 16. Arena Vision leader → Multimodal 18. Pricing: $2/$12 per M tokens. 1M context window (beta).