aiXamine - Fairness & Bias: leaderboard
aiXamine Fairness & Bias service score: consistency and impartiality across social groups on Adult, gender-care, preference and BBQ tests (aiXamine, arXiv 2608.20554v1, Appendix Table 7). Normalized 0-100 score aggregated from the service's tests; higher is better. 42 models evaluated under identical black-box conditions via OpenRouter or containerized vLLM.
Metric: Service score (0-100; normalized pass rate). Source: arxiv.org. Saturation forecast: Around January 2028. 42 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B | 79 | #212 |
| 2 | Gemma 4 26B A4B | 77.5 | #360 |
| 3 | Gemini 3 Flash (Preview) | 76.7 | #90 |
| 4 | GPT-OSS-120B | 72.9 | #346 |
| 5 | Phi-4 | 72 | #707 |
| 6 | Qwen 3.5 27B | 71.2 | #192 |
| 7 | GPT-5 Mini | 68.8 | #193 |
| 8 | GPT-OSS-20B | 67.6 | #523 |
| 9 | GPT-4o | 67.5 | #357 |
| 10 | Gemini 2.0 Flash (001) | 66.9 | #365 |
| 11 | Gemma 4 E4B | 65.8 | #762 |
| 12 | Llama 3.1 8B | 65.7 | #1130 |
| 13 | Gemma 2 27B | 64.7 | #803 |
| 14 | Qwen 3.5 122B A10B | 64.3 | #184 |
| 15 | DeepSeek R1 Distill Qwen 14B | 64 | #860 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aixamine-fairness-bias · How It Works · Data refreshed daily, snapshot 2026-11-05.