OpenVLM MMT-Bench - Weapon Recognition: leaderboard
Metric: Score (%). Source: huggingface.co. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 100 |
| 2 | GPT-4.1 Mini | 95 |
| 3 | InternVL3-78B | 95 |
| 4 | InternVL3-38B | 95 |
| 5 | Gemma 3 27B | 95 |
| 6 | Grok 2 (1212) | 95 |
| 7 | Gemma 3 12B | 95 |
| 8 | Llama 3.2 90B Vision Instruct | 95 |
| 9 | Qwen 2 VL 72B | 90 |
| 10 | Qwen 2 VL 2B | 90 |
| 11 | Gemini 1.5 Pro | 90 |
| 12 | InternVL3-8B | 85 |
| 13 | Qwen 2 VL 7B | 85 |
| 14 | Aquila-VL-2B | 85 |
| 15 | Claude 3.7 Sonnet | 85 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-weapon-recognition · How It Works · Data refreshed daily, snapshot 2026-09-05.