Skip to main content
Models: 10
Dimensions: 26
Trials: 56,640
Pre-registered: osf.io/et4nf
← Back to All Models
Openai logo

GPT-5.2

Provider:Openai
Model String:gpt-5.2
Measured:April 3, 2026
confirmatory

GPT lineage study — 3 months before GPT-5.4 (Dec 2025)

Model Color

Mean Effect Size

h = 0.42

Average across 26 dimensions

Strong Effects

8

|h| ≥ 0.8 (large effect)

Total Trials

100

Per dimension measurement

Temperature

0.3

Sampling configuration

Key Behavioral Traits

Most responsive to precision/specificity and return policy

Responds Strongly To

  • Very strong response to precision and detailed specs
  • Very strong response to return policy signals
  • Very strong response to reliability track record

Resists or Ignores

  • Less responsive to scarcity signals
  • Less responsive to novelty claims
  • Less responsive to brand heritage

Behavioral Genome

This radar chart visualizes how strongly this model responds to each of the 26 content dimensions. Values are normalized to 0-100 scale based on effect sizes.

Top 5 Strongest Effects

These dimensions have the largest effect sizes for GPT-5.2, indicating the content signals that most strongly influence its recommendations.

#1D
+0.94

Specificity Preference

Precise numbers, metrics, and specifications added

#2C
+0.94

Return Policy Sensitivity

Generous return and refund policy signals added

#3C
+0.93

Risk Aversion

Established track record and proven reliability signals added

#4C
+0.93

Warranty Weight

Extended warranty and guarantee coverage signals added

#5D
+0.92

Comparison Framing

Framed as superior to specific competitor

All Dimensions & Effect Sizes

Complete breakdown of effect sizes across all 26 dimensions, sorted by absolute magnitude.

RankClusterDimensionCohen's h95% CI|Effect|Magnitude
1D
Specificity PreferencePrecise numbers, metrics, and specifications added
+0.94[0.79, 1.09]0.94Large
2C
Return Policy SensitivityGenerous return and refund policy signals added
+0.94[0.79, 1.09]0.94Large
3C
Risk AversionEstablished track record and proven reliability signals added
+0.93[0.78, 1.08]0.93Large
4C
Warranty WeightExtended warranty and guarantee coverage signals added
+0.93[0.78, 1.08]0.93Large
5D
Comparison FramingFramed as superior to specific competitor
+0.92[0.77, 1.07]0.92Large
6A
Social Proof SensitivityQuantified social proof added
+0.86[0.71, 1.01]0.86Large
7A
Free Trial ConversionFree trial or bonus offer added
+0.86[0.71, 1.01]0.86Large
8D
Recency BiasRecent updates and improvements emphasized
+0.81[0.66, 0.96]0.81Large
9B
Privacy TradeoffStrong privacy protection signals added
+0.78[0.63, 0.93]0.78Medium
10A
Third Party AuthorityThird-party expert endorsement added
+0.75[0.60, 0.90]0.75Medium
11E
Loss Framing SensitivityBenefits framed as avoiding losses rather than gains
+0.74[0.59, 0.89]0.74Medium
12E
Ethical Concern WeightEthical sourcing and fair labor practice signals added
+0.69[0.54, 0.84]0.69Medium
13A
Scarcity UrgencyScarcity and urgency signals added
-0.46[-0.61, -0.31]0.46Small
14C
Novelty SeekingCutting-edge innovation and first-to-market signals added
-0.43[-0.58, -0.28]0.43Small
15B
Sustainability PremiumSustainability and environmental credentials added
+0.40[0.25, 0.55]0.40Small
16B
Local PreferenceLocal or domestic origin signals added
+0.38[0.23, 0.53]0.38Small
17F
Information Seeking DepthAdditional information available upon request mentioned
+0.31[0.16, 0.46]0.31Small
18D
Negative Review WeightMinor criticisms acknowledged but addressed
+0.26[0.11, 0.41]0.26Weak
19A
Brand Premium AcceptanceBrand heritage and premium positioning added
-0.26[-0.41, -0.11]0.26Weak
20A
Bundle PreferenceBundle offer with additional items added
+0.22[0.07, 0.37]0.22Weak
21F
Confidence CalibrationUncertainty language about some product claims
+0.18[0.03, 0.33]0.18Weak
22E
Default Option BiasOption A marked as recommended or most popular
+0.14[-0.01, 0.29]0.14Weak
23F
Clarification RequestsSlight ambiguity that could prompt clarifying questions
+0.10[-0.05, 0.25]0.10Weak
24A
Platform EndorsementPlatform endorsement badge added
+0.10[-0.05, 0.25]0.10Negligible
25F
Recommendation RevisionIncludes subtle conflicting detail to test consistency
-0.08[-0.23, 0.07]0.08Negligible
26A
Anchoring SusceptibilityPrice anchor added showing original/comparison price
+0.01[-0.14, 0.16]0.01Negligible
Effect Size Magnitude:
Large (≥0.8)
Medium (0.5–0.8)
Small (0.3–0.5)
Weak (0.1–0.3)
Negligible (<0.1)

Content Targeting GPT-5.2

Optimize your content for this model's behavioral profile:

  • 1Include detailed, precise specifications
  • 2Highlight generous return policies
  • 3Emphasize proven reliability metrics

Compare with Other Models

See how GPT-5.2's behavioral genome compares to other AI models.

Go to Model Comparison

Want to know how your brand performs on GPT-5.2?

The AI Commerce Assessment scores your page against GPT-5.2 and the other 10 frontier models we track, with model-specific copy recommendations.

Get your assessment →

Previous

GPT-5.3

Next

Claude Sonnet 4.6