Evaluation profile
OpenAI GPT-5.4 First-Person Fairness
1sub-evals
0.055%total index weight
1components
Within-component eval weight: Fairness 0.55%.
Model score (lower is better)Predicted score
About this eval
OpenAI GPT-5.4 First-Person Fairness
Included in the behavior ranking.
Sub-evals
| Measure | Component | Direction | Total index weight | Within-component weight |
|---|---|---|---|---|
| harm_overallopenai-gpt54/first-person-fairness.csv:harm_overallMeasures harm overall in the cited first-party or paper evaluation. | fairness_nondiscrimination:1.000openai-gpt54-system-card | Lower is better | 0.055% | Fairness 0.55% |
harm_overall
Measures harm overall in the cited first-party or paper evaluation.
| Rank | Model | Value | Relative performance | Provenance |
|---|---|---|---|---|
| 1 | gpt-5.4 | 0.0088 | official | |
| 2 | gpt-5.2-thinking | 0.00997 | official | |
| 3 | gpt-5.1 | 0.0128 | official |