Spatial Reasoning on MindCube tiny (Overall, Around, Among, Rotation)
81.6Overall ScoreGemini-3-Pro
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini-3-ProAlloSpatial=Integrated, Setting=training-free2026.06 | 81.6 | 86 | 75.8 | 93.5 | |
| Gemini-3-ProAlloSpatial=None, Setting=training-free2026.06 | 75.1 | 77.2 | 68.2 | 93 | |
| InternVL3-8B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=8B2026.06 | 73.7 | 86 | 77.2 | 38.5 | |
| InternVL3-8B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=8B2026.06 | 68.5 | 81.5 | 70.8 | 35.5 | |
| InternVL3-8B + TextSupervision=Textual, Model Family=InternVL3, Model Size=8B2026.06 | 67.5 | 82.5 | 68.8 | 33.5 | |
| LLaVA-NeXT-Video-7B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.06 | 64.9 | 82.3 | 61.8 | 39.5 | |
| Claude-4.6-OpusAlloSpatial=Integrated, Setting=training-free2026.06 | 62.9 | 82.4 | 60.8 | 45 | |
| Qwen2.5-VL-7B + SpatialSVSupervision=SpatialSV, Model Family=Qwen2.5-VL, Model Size=7B2026.06 | 62.8 | 81.8 | 58.5 | 37.5 | |
| InternVL3-2B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=2B2026.06 | 62.4 | 82.8 | 57.2 | 37.5 | |
| Qwen2.5-VL-3B + SpatialSVSupervision=SpatialSV, Model Family=Qwen2.5-VL, Model Size=3B2026.06 | 62.3 | 84.5 | 56.2 | 36 | |
| InternVL3-2B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=2B2026.06 | 60.6 | 79.5 | 56.5 | 35 | |
| InternVL3-1B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=1B2026.06 | 60.4 | 78.5 | 56.8 | 35 | |
| LLaVA-NeXT-Video-7B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.06 | 60.2 | 79.5 | 55.2 | 36.5 | |
| Qwen2.5-VL-3B + DistillationSupervision=Feature Distillation, Model Family=Qwen2.5-VL, Model Size=3B2026.06 | 58.6 | 81.3 | 51.5 | 34.5 | |
| InternVL3-2B + TextSupervision=Textual, Model Family=InternVL3, Model Size=2B2026.06 | 58.5 | 78 | 54.3 | 32 | |
| InternVL3-1B + TextSupervision=Textual, Model Family=InternVL3, Model Size=1B2026.06 | 58.4 | 80 | 52.3 | 33.5 | |
| Qwen2.5-VL-7B + DistillationSupervision=Feature Distillation, Model Family=Qwen2.5-VL, Model Size=7B2026.06 | 58.3 | 76.5 | 53.8 | 35 | |
| InternVL3-1B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=1B2026.06 | 58.3 | 77 | 53.5 | 35.5 | |
| LLaVA-NeXT-Video-7B + TextSupervision=Textual, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.06 | 57.9 | 78.8 | 52.3 | 33 | |
| Qwen2.5-VL-7B + TextSupervision=Textual, Model Family=Qwen2.5-VL, Model Size=7B2026.06 | 55.9 | 73.5 | 51 | 35.5 | |
| Qwen2.5-VL-3B + TextSupervision=Textual, Model Family=Qwen2.5-VL, Model Size=3B2026.06 | 55.3 | 74.8 | 49.7 | 33.5 | |
| LLaVA-OneVision-1.5-8B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.06 | 55.2 | 60.3 | 57.5 | 38 | |
| GPT-5.2AlloSpatial=Integrated, Setting=training-free2026.06 | 54.6 | 60.4 | 47.7 | 68 | |
| LLaVA-OneVision-1.5-8B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.06 | 51.8 | 58.3 | 52.7 | 36.5 | |
| LLaVA-OneVision-1.5-4B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.06 | 51.2 | 51.5 | 56.5 | 34.5 | |
| GPT-5.2AlloSpatial=None, Setting=training-free2026.06 | 49.9 | 62.4 | 45.2 | 48.5 | |
| LLaVA-OneVision-1.5-8B + TextSupervision=Textual, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.06 | 49.1 | 55.5 | 49.5 | 35 | |
| Claude-4.6-OpusAlloSpatial=None, Setting=training-free2026.06 | 48.5 | 58.8 | 50.7 | 29 | |
| LLaVA-OneVision-1.5-4B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.06 | 45.4 | 42.3 | 51.2 | 34.5 | |
| LLaVA-OneVision-1.5-4B + TextSupervision=Textual, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.06 | 43.1 | 38.3 | 49 | 35 | |
| LLaVA-NeXT-Video-7BSupervision=None, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.06 | 38.1 | 34.5 | 41.7 | 34.5 | |
| InternVL3-8BSupervision=None, Model Family=InternVL3, Model Size=8B2026.06 | 36.5 | 33.3 | 39.3 | 34.5 | |
| InternVL3-1BSupervision=None, Model Family=InternVL3, Model Size=1B2026.06 | 34.6 | 32 | 37.3 | 31.5 | |
| Qwen2.5-VL-3BSupervision=None, Model Family=Qwen2.5-VL, Model Size=3B2026.06 | 34.5 | 32.3 | 36 | 34.5 | |
| Chance Level (Frequency)Type=Baseline2026.06 | 34.3 | 33.3 | 34.8 | 34.5 | |
| LLaVA-OneVision-1.5-4BSupervision=None, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.06 | 32.8 | 32.3 | 35.2 | 26.5 | |
| LLaVA-OneVision-1.5-8BSupervision=None, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.06 | 32 | 32.5 | 31 | 33.5 | |
| InternVL3-2BSupervision=None, Model Family=InternVL3, Model Size=2B2026.06 | 30.1 | 24.8 | 33.3 | 31 | |
| Qwen2.5-VL-7BSupervision=None, Model Family=Qwen2.5-VL, Model Size=7B2026.06 | 29.6 | 26.8 | 29.7 | 35 | |
| Chance Level (Random)Type=Baseline2026.06 | 24.6 | 23.8 | 24.5 | 26.5 |