Hire Fairly Resume Audit - Gender: leaderboard
The same paired callback audit with 48,048 male- and female-coded resume pairs per model (Grok-4.1-fast has no gender run); the score is the absolute callback gap in percentage points.
Metric: Absolute gender selection gap between paired resumes (percentage points). Source: arxiv.org. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 0.03 | #453 |
| 2 | GPT-5.4 Mini | 0.11 | #389 |
| 3 | Gemma 4 31B (IT) | 0.18 | #342 |
| 4 | Claude Haiku 4.5 | 0.41 | #568 |
| 5 | Llama 3.3 70B Instruct | 0.53 | #1357 |
| 6 | Qwen 3.5 397B A17B | 0.63 | #246 |
| 7 | GLM-5.1 | 0.67 | #179 |
| 8 | GPT-OSS-120B | 0.68 | #700 |
| 9 | Llama 3.1 8B Instruct | 0.84 | #4246 |
| 10 | GPT-4o Mini | 1.85 | #1623 |
| 11 | GPT-3.5 Turbo | 1.92 | #2753 |
| 12 | DeepSeek V3.1 | 2.12 | #497 |
| 13 | Claude 3 Haiku | 5.8 | #2609 |
Interactive version: theaggregate.ai/benchmark?slug=hire-fairly-resume-audit-gender · How It Works · Data refreshed daily, snapshot 2026-11-13.