MM-JudgeBias - Detail Description: 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-Conformity from 0 to 1, the stability of the score when a detailed caption of the image is appended to the query; higher is better.

Metric: Bias-Conformity (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 score invariance (1 minus the absolute score change over its largest possible value) when a detailed caption of the image is appended to the query; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 26 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 ProHigh0.936#146
2Gemini 2.5 Pro0.935#280
3LLaVA-Critic-72B0.926
4Claude Opus 4.5Thinking0.917#165
5GPT-5.1High0.911#248
6Gemini 2.5 Flash0.905#466
7GPT-5 MiniHigh0.898#373
8Gemini 2.0 Flash Lite0.897#1064
9Qwen 3 VL 30B A3B (Thinking)0.886#727
10O4 MiniHigh0.886#372
11Claude Sonnet 4.5Thinking0.884#267
12Qwen 3 VL 8B (Thinking)0.883#924
13GPT-4.1 Mini0.882#799
14O3High0.879#242
15Claude Haiku 4.5Thinking0.87#587

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=mm-judgebias-detail-description · How It Works · Data refreshed daily, snapshot 2026-11-22.