OpenVLM MMT-Bench - 3D Indoor Recognition: leaderboard
Metric: Score (%). Source: huggingface.co. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.7 Sonnet | 60 |
| 2 | GPT-4.1 Mini | 50 |
| 3 | Claude 3.5 Sonnet | 45 |
| 4 | Gemini 1.5 Flash | 45 |
| 5 | Gemini 1.5 Pro | 40 |
| 6 | Gemma 3 12B | 40 |
| 7 | Llama 3.2 11B Instruct | 35 |
| 8 | GPT-4.1 (2025-04-14) | 35 |
| 9 | InternVL3-78B | 30 |
| 10 | Gemma 3 27B | 30 |
| 11 | InternVL3-38B | 25 |
| 12 | Qwen 2 VL 2B | 25 |
| 13 | Pixtral-12B | 25 |
| 14 | Llama 3.2 90B Vision Instruct | 25 |
| 15 | InternVL2-8B | 20 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-3d-indoor-recognition · How It Works · Data refreshed daily, snapshot 2026-09-05.