MoralMetric

LLM Bias Analysis & Comparison

Models
11
Categories
2
Total Axes
7
Avg Bias
19.2%
Least Biased
Claude 4.5 Opus2.7%
Most Biased
Grok 4.1 Fast53.2%
Updated: February 1, 2026 at 02:43 PM

Political Bias

Tests for implicit political bias in LLM decision-making through counterfactual A/B testing.

0
A
Claude 4.5 Opus
0
A
Claude 4.5 Sonnet
0
G
Gemini 3 Flash
11Right
O
GPT OSS 120B
13Left
O
GPT-5.2
15Left
M
Kimi K2.5
19Left
X
Grok 4
25Left
Z
GLM 4.7
26Left
D
DeepSeek V3.2
59Right
G
Gemini 3 Pro
100Right
X
Grok 4.1 Fast
Models Ranked by Bias Score (Lower is Better)
anthropic
google
openai
moonshotai
x-ai
z-ai
deepseek
#1
0.0%
Overall Bias0.0%
Axes Tested1
Preference OrderLeftRight

Full Preference Ranking

1Left
2Right

Axis-by-Axis Breakdown

Left Vs RightMinimal
LeftNeutralRight
0.0%
100%50%050%100%