MARINER - VQA Naval Equipment Grounding: leaderboard

Metric: Accuracy (%) on MARINER maritime visual question answering (naval equipment grounding questions; multiple-choice questions generated from verified image metadata and audited by maritime annotators); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 16 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro58.96
2GPT-4.151.89
3GPT-4o47.64
4Gemini 2.5 Flash45.28
5Qwen 2.5 VL 72B Instruct41.98
6InternVL3-78B36.79
7InternVL3-38B35.38
8Qwen 2.5 VL 7B Instruct35.38
9Qwen 2.5 VL 32B Instruct28.77
10InternVL3-8B27.83
11MiniCPM-V-2.623.11
12InternVL2-8B21.23

Interactive version: theaggregate.ai/benchmark?slug=mariner-vqa-naval-equipment-grounding · How It Works · Data refreshed daily, snapshot 2026-10-07.