GenderBench - Jobs (Lum et al.) Gender Imbalance: leaderboard
In the occupation writing task (Lum et al. jobs), the number is how far the share of male protagonists is from 0.5, regardless of stereotype.
Metric: Distance of the male-protagonist rate from 0.5 (occupations) (0-1 normalized, lower is better). Source: genderbench.readthedocs.io. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 8B Instruct | 0.08 | #4246 |
| 2 | Mistral 7B Instruct (v0.3) | 0.1 | #6653 |
| 3 | Gemini 2.0 Flash Lite | 0.11 | #1032 |
| 4 | Phi-4 | 0.12 | #1913 |
| 5 | Mistral Small 3 | 0.15 | #1777 |
| 6 | Llama 3.3 70B Instruct | 0.19 | #1357 |
| 7 | Gemma 2 9B (IT) | 0.19 | #2262 |
| 8 | GPT-4o | 0.2 | #741 |
| 9 | Gemini 2.0 Flash | 0.2 | #704 |
| 10 | Gemma 2 27B (IT) | 0.21 | #1782 |
| 11 | GPT-4o Mini | 0.21 | #1623 |
| 12 | Claude 3.5 Haiku | 0.23 | #1553 |
Interactive version: theaggregate.ai/benchmark?slug=genderbench-jobs-lum-et-al-gender-imbalance · How It Works · Data refreshed daily, snapshot 2026-11-13.