MARINER - VQA Perspective Estimation: leaderboard

Maritime visual question answering on MARINER (2026), with multiple-choice questions built from verified image metadata and audited by maritime annotators. Accuracy in percent on vessel perspective questions; higher is better.

Metric: Accuracy (%) on MARINER maritime visual question answering (vessel perspective questions; multiple-choice questions generated from verified image metadata and audited by maritime annotators); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 16 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-4.172.73#505
2Gemini 2.5 Pro68.53#276
3GPT-4o62.7#751
4Gemini 2.5 Flash59.67#463
5LLaVA-OneVision-72B57.34#1771
6InternVL3-78B56.41#759
7Qwen 2.5 VL 72B Instruct56.18#841
8InternVL3-38B55.94#939
9Qwen 2.5 VL 7B Instruct51.52#1902
10InternVL3-8B50.58#1782
11Qwen 2.5 VL 32B Instruct49.65#1062
12MiniCPM-V-2.644.29#3267
13LLaVA-OneVision-7B41.96#2981
14InternVL2-8B41.72#3227
15Qwen 2.5 VL 3B Instruct39.16#2713

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=mariner-vqa-perspective-estimation · How It Works · Data refreshed daily, snapshot 2026-11-21.