MARINER - VQA: leaderboard
Metric: Accuracy (%) on MARINER maritime visual question answering (mean of the eight task types in perception, space and reasoning; multiple-choice questions generated from verified image metadata and audited by maritime annotators); higher is better. Source: arxiv.org. Saturation forecast: Around April 2027. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 73.97 |
| 2 | GPT-4.1 | 73.77 |
| 3 | InternVL3-38B | 69.29 |
| 4 | InternVL3-78B | 69.05 |
| 5 | GPT-4o | 68.91 |
| 6 | Qwen 2.5 VL 72B Instruct | 68.79 |
| 7 | Qwen 2.5 VL 7B Instruct | 65.63 |
| 8 | InternVL3-8B | 64.74 |
| 9 | Gemini 2.5 Flash | 64.52 |
| 10 | Qwen 2.5 VL 32B Instruct | 63.85 |
| 11 | MiniCPM-V-2.6 | 61.57 |
| 12 | InternVL2-8B | 60.08 |
Interactive version: theaggregate.ai/benchmark?slug=mariner-vqa · How It Works · Data refreshed daily, snapshot 2026-10-07.