MM-JudgeBias - Text Dominance: leaderboard
MM-JudgeBias (2026) tests multimodal LLM judges for compositional bias: the judge scores image, query and response triplets from 1 to 10 before and after a controlled perturbation. Bias-Deviation from 0 to 1, the normalized score drop when the image is replaced with a black image; higher is better.
Metric: Bias-Deviation (0 to 1) on MM-JudgeBias (about 200 image-query-response triplets per bias type, 1,804 in all, each judged on a 1-10 scale before and after a controlled perturbation; mean of three runs): mean drop of the judge's score, normalized by the largest possible drop, when the image is replaced with a black image; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 26 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3 Pro | High | 0.912 | #145 |
| 2 | Gemini 2.5 Pro | 0.751 | #276 | |
| 3 | Claude Opus 4.5 | Thinking | 0.68 | #160 |
| 4 | Claude Sonnet 4.5 | Thinking | 0.556 | #261 |
| 5 | Gemini 2.5 Flash | 0.423 | #463 | |
| 6 | Gemini 2.5 Flash Lite | Thinking | 0.391 | #964 |
| 7 | Qwen 3 VL 8B Instruct | 0.336 | #1029 | |
| 8 | Qwen 3 VL 30B A3B (Thinking) | 0.293 | #719 | |
| 9 | Claude Haiku 4.5 | Thinking | 0.291 | #583 |
| 10 | O3 | High | 0.276 | #237 |
| 11 | LLaVA-Critic-7B | 0.238 | ||
| 12 | Qwen 3 VL 30B A3B Instruct | 0.237 | #895 | |
| 13 | GPT-5.1 | High | 0.192 | #240 |
| 14 | Qwen 3 VL 8B (Thinking) | 0.177 | #910 | |
| 15 | Prometheus-Vision-7B | 0.167 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=mm-judgebias-text-dominance · How It Works · Data refreshed daily, snapshot 2026-11-21.