Mono3DRefer (M3Sun) - mIoU: leaderboard

Monocular 3D visual grounding on the Mono3DRefer test set: a general-purpose VLM predicts the referred object's 3D bounding box from an image and language. This is the M3Sun paper's direct VLM baseline protocol; task-trained vision models and the multi-tool M3SunAgent have separate protocols. Checkpoint names and unspecified API effort are retained as printed. Score is mean 3D intersection over union as a percentage.

Metric: Mean 3D intersection over union (%). Source: arxiv.org. Saturation forecast: Around 2028. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-57.53#185
2O35.93#242
3Gemini 2.5 Pro2.37#280
4MiMo-VL-7B1.71#1457
5Gemini 2.5 Flash1.29#466
6Qwen 2.5 VL 72B0.89#782

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=mono3drefer-m3sun-miou · How It Works · Data refreshed daily, snapshot 2026-11-22.