MORPHOGEN - Arabic Gender IoU: leaderboard
Metric: Gender IoU (%): per-sentence ratio of correctly transformed gendered terms to the union of gendered and wrongly changed terms, averaged over sentences, in the GENFORM task of MORPHOGEN: rewrite a first-person Arabic sentence for a speaker of the opposite grammatical gender while keeping its meaning (both directions, human-corrected references), zero-shot with the model chat template, temperature 0.1; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o Mini | 71.02 |
| 2 | Gemma 3 27B | 70.33 |
| 3 | Gemma 3 12B | 62.76 |
| 4 | Llama 3.3 70B | 59.16 |
| 5 | Phi-4 | 57.08 |
| 6 | Qwen 3 14B | 51.83 |
| 7 | Qwen 3 32B | 50.69 |
| 8 | Gemma 2 9B | 46.45 |
| 9 | Qwen 3 8B | 45.89 |
| 10 | Gemma 3 4B | 45.68 |
| 11 | Llama 3.1 8B | 43.51 |
| 12 | Qwen 3 4B | 34.34 |
| 13 | Llama 3.2 3B | 18.31 |
| 14 | Gemma 2 2B | 14.73 |
| 15 | Qwen 2.5 0.5B | 4.14 |
Interactive version: theaggregate.ai/benchmark?slug=morphogen-arabic-gender-iou · How It Works · Data refreshed daily, snapshot 2026-10-07.