| [ |
| { |
| "model_name":"Human", |
| "score":79.2, |
| "VSI [66]":"79.2", |
| "SITE [57]":"67.5", |
| "MMSI [68]":"97.2", |
| "OmniSpatial [23]":"92.63", |
| "MindCube ∗ [69]":"94.55", |
| "STARE [32]":"96.50", |
| "CoreCognition [33]":"86.98", |
| "SpatialViz [55]":"82.46", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Qwen3-8B-Instruct [65]", |
| "score":57.9, |
| "VSI [66]":"57.90", |
| "SITE [57]":"45.83", |
| "MMSI [68]":"31.10", |
| "OmniSpatial [23]":"45.73", |
| "MindCube ∗ [69]":"29.42", |
| "STARE [32]":"39.76", |
| "CoreCognition [33]":"69.67", |
| "SpatialViz [55]":"17.54 †", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"† indicates cases where generations were truncated due to overlong chains of thought, yielding no final answer; such instances are counted as incorrect, which depresses the score." |
| }, |
| { |
| "model_name":"InternVL3.5-8B [56]", |
| "score":56.05, |
| "VSI [66]":"56.05", |
| "SITE [57]":"43.79", |
| "MMSI [68]":"27.30", |
| "OmniSpatial [23]":"46.71", |
| "MindCube ∗ [69]":"42.50", |
| "STARE [32]":"40.18", |
| "CoreCognition [33]":"66.40", |
| "SpatialViz [55]":"23.98", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"GPT-5-2025-08-07 [45]", |
| "score":55.03, |
| "VSI [66]":"55.03", |
| "SITE [57]":"61.88", |
| "MMSI [68]":"41.80", |
| "OmniSpatial [23]":"59.90", |
| "MindCube ∗ [69]":"56.30", |
| "STARE [32]":"54.59", |
| "CoreCognition [33]":"84.37", |
| "SpatialViz [55]":"51.27", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Gemini-2.5-pro-2025-06 [52]", |
| "score":53.57, |
| "VSI [66]":"53.57", |
| "SITE [57]":"57.06", |
| "MMSI [68]":"38.00", |
| "OmniSpatial [23]":"55.38", |
| "MindCube ∗ [69]":"57.60", |
| "STARE [32]":"49.14", |
| "CoreCognition [33]":"76.70", |
| "SpatialViz [55]":"42.71", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Seed-1.6-2025-06-15 [51]", |
| "score":49.91, |
| "VSI [66]":"49.91", |
| "SITE [57]":"54.61", |
| "MMSI [68]":"38.30", |
| "OmniSpatial [23]":"49.32", |
| "MindCube ∗ [69]":"48.75", |
| "STARE [32]":"46.06", |
| "CoreCognition [33]":"77.17", |
| "SpatialViz [55]":"34.58", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"GPT-5-mini-2025-08-07 [45]", |
| "score":48.67, |
| "VSI [66]":"48.67", |
| "SITE [57]":"52.47", |
| "MMSI [68]":"34.10", |
| "OmniSpatial [23]":"55.52", |
| "MindCube ∗ [69]":"56.69", |
| "STARE [32]":"52.51", |
| "CoreCognition [33]":"77.77", |
| "SpatialViz [55]":"44.66", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Grok-4-2025-07-09 [62]", |
| "score":47.92, |
| "VSI [66]":"47.92", |
| "SITE [57]":"47.01", |
| "MMSI [68]":"37.80", |
| "OmniSpatial [23]":"46.84", |
| "MindCube ∗ [69]":"63.56", |
| "STARE [32]":"26.90", |
| "CoreCognition [33]":"79.27", |
| "SpatialViz [55]":"19.40 †", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"† indicates cases where generations were truncated due to overlong chains of thought, yielding no final answer; such instances are counted as incorrect, which depresses the score." |
| }, |
| { |
| "model_name":"InternVL3-78B [79]", |
| "score":47.55, |
| "VSI [66]":"47.55", |
| "SITE [57]":"52.72", |
| "MMSI [68]":"30.50", |
| "OmniSpatial [23]":"50.95", |
| "MindCube ∗ [69]":"49.52", |
| "STARE [32]":"42.00", |
| "CoreCognition [33]":"71.16", |
| "SpatialViz [55]":"31.10", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"GPT-5-nano-2025-08-07 [45]", |
| "score":43.22, |
| "VSI [66]":"43.22", |
| "SITE [57]":"35.81", |
| "MMSI [68]":"28.90", |
| "OmniSpatial [23]":"47.81", |
| "MindCube ∗ [69]":"41.48", |
| "STARE [32]":"46.05", |
| "CoreCognition [33]":"67.92", |
| "SpatialViz [55]":"35.59", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"InternVL3-8B [79]", |
| "score":42.14, |
| "VSI [66]":"42.14", |
| "SITE [57]":"41.15", |
| "MMSI [68]":"28.00", |
| "OmniSpatial [23]":"46.25", |
| "MindCube ∗ [69]":"41.54", |
| "STARE [32]":"41.36", |
| "CoreCognition [33]":"60.92", |
| "SpatialViz [55]":"30.00", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Qwen2.5-VL-72B-Instruct [1]", |
| "score":35.77, |
| "VSI [66]":"35.77", |
| "SITE [57]":"47.41", |
| "MMSI [68]":"32.50", |
| "OmniSpatial [23]":"47.81", |
| "MindCube ∗ [69]":"42.40", |
| "STARE [32]":"38.37", |
| "CoreCognition [33]":"69.22", |
| "SpatialViz [55]":"32.54", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Random Choice", |
| "score":34.0, |
| "VSI [66]":"34.00", |
| "SITE [57]":"0.0", |
| "MMSI [68]":"25.00", |
| "OmniSpatial [23]":"24.98", |
| "MindCube ∗ [69]":"32.35", |
| "STARE [32]":"34.80", |
| "CoreCognition [33]":"33.93", |
| "SpatialViz [55]":"25.08", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"VSI random choice here is chance level(Frequency)." |
| }, |
| { |
| "model_name":"Qwen2.5-VL-7B-Instruct [1]", |
| "score":32.3, |
| "VSI [66]":"32.30", |
| "SITE [57]":"37.64", |
| "MMSI [68]":"26.80", |
| "OmniSpatial [23]":"39.07", |
| "MindCube ∗ [69]":"36.05", |
| "STARE [32]":"35.03", |
| "CoreCognition [33]":"62.16", |
| "SpatialViz [55]":"26.78", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| }, |
| { |
| "model_name":"Qwen2.5-VL-3B-Instruct [1]", |
| "score":27.0, |
| "VSI [66]":"27.00", |
| "SITE [57]":"33.14", |
| "MMSI [68]":"28.60", |
| "OmniSpatial [23]":"42.47", |
| "MindCube ∗ [69]":"37.60", |
| "STARE [32]":"37.83", |
| "CoreCognition [33]":"60.19", |
| "SpatialViz [55]":"21.86", |
| "source_title":"EASI", |
| "source_url":"https:\/\/arxiv.org\/pdf\/2508.13142.pdf", |
| "notes":"None" |
| } |
| ] |
