List your company

Spatial Reasoning (Granular Metrics) on MMSI-Bench

49.5Average AccuracyMSSR
23.530.253743.75Oct 19, 2025Nov 28, 2025Jan 7, 2026Feb 17, 2026Mar 29, 2026May 8, 2026Jun 18, 2026

Evaluation results

MethodLinks
MSSRModel Category=Agentic, Backbone=GPT-4o, Evaluation Protocol=zero-shot2025.1049.5——————————5050.647.1——
Gemini-3-FlashWorkflow Mode=Forced Tool-Use (zero-shot), Type=Proprietary2026.0649.5——————————50.350.4—51.543.4
S-AgentModel Category=Ours2026.0646.446.243.6374343.563.957.840.94648.744.4————
Gemini-3-FlashWorkflow Mode=Direct Answer, Type=Proprietary2026.0645.4——————————4645.6—45.444
Gemini 3 ProModel Category=Proprietary Models2026.0645.247.348.9424337.660.264.139.441.947.437.9————
GPT-5Model Category=Proprietary2026.0341.94335.132.148.842.451.860.936.432.436.842————
GPT-5.4Model Category=Proprietary Models2026.0641.941.93335.849.842.468.754.737.428.340.836.4————
o3Model Category=Proprietary2026.034145.239.43744.247.162.654.728.831.132.934.9————
o3Model Category=Proprietary LLM2025.1041——————————34.945.836.4——
GPT-4.5Model Category=Proprietary2026.0340.334.429.839.551.247.155.439.133.341.940.836.4————
LEOModel Category=Specialist2025.1039.3——————————32.342.338.6——
Astra Qwen3-VL-8B-InstructWorkflow Mode=Agentic Tool-Use, Type=Open-source2026.0638.8——————————33.642.3—4132.1
Seed 1.6Model Category=Open-weight General Models2026.0638.536.636.232.132.642.446.948.43331.142.140.4————
SN-SI-1.1-Qwen3VL-8BModel Category=Open-weight Spatial Models2026.0638.144.138.333.365.138.85948.424.229.734.222.2————
Gemini 2.5 ProModel Category=Proprietary Models2026.063838.73440.744.238.84162.530.339.22533.3————
SN-U1-8B-MoTModel Category=Open-weight General Models2026.063846.241.529.658.138.863.943.821.225.731.626.8————
Grok 4Model Category=Proprietary Models2026.0637.836.635.139.534.945.950.621.922.740.543.438.4————
pySpatialModel Category=Baseline, Base Model=GPT-4o2026.0337.351.628.727.220.941.238.646.939.44638.236.4————
Gemini 2.5 ProModel Category=Proprietary LLM2025.1037——————————34.338.536.1——
Gemini-2.5-ProModel Category=Proprietary2026.0336.939.731.939.545.335.243.351.521.236.430.234.3————
Gemini-2.5-ProWorkflow Mode=Direct Answer, Type=Proprietary2026.0636.9——————————34.339—36.233.3
Qwen3.5-9BModel Category=Open-weight General Models2026.0636.534.436.234.639.538.854.256.328.836.526.328.8————
Gemini 2.5 FlashModel Category=Proprietary LLM2025.1035——————————30.337.433.9——
GLM-4.5VWorkflow Mode=Direct Answer, Type=Proprietary2026.0633.8——————————30.335.6—36.929.3
SN-SI-1.1-Qwen2.5VL-7BModel Category=Open-weight Spatial Models2026.0632.851.629.832.15029.442.237.528.82334.218.7————
VST-7B-SFTModel Category=Open-weight Spatial Models2026.0632.539.836.235.837.229.433.729.74736.535.518.2————
VLM-3RModel Category=3D-VLM2025.1032——————————30.33331.4——
Qwen3-VL-4BModel Category=Open-source LLM2025.1031.1——————————27.832.830.3——
Qwen3-VL-8BModel Category=Open-source LLM2025.1031.1——————————30.332.628.9——
Qwen3-VL-8B-InstructModel Category=Open-weight General Models2026.0631.12837.232.131.435.338.537.515.22728.929.8————
Bagel-7B-MoTWorkflow Mode=Direct Answer, Type=Open-source2026.0631——————————30.833.5—27.725.3
GPT-4.1Model Category=Proprietary2026.0330.936.626.627.229.136.527.737.524.236.532.928.8————
Qwen2.5-VL-72BModel Category=Open-source2026.0330.725.83434.623.334.136.145.327.32730.327.3————
Qwen2.5-VL-72BModel Category=Open-source LLM2025.1030.7——————————27.331.232.1——
Qwen3-VL-30B-InstructWorkflow Mode=Direct Answer, Type=Open-source2026.0630.6——————————29.131.2—35.825.9
GPT-4oModel Category=Proprietary2026.0330.334.424.523.519.837.627.732.831.835.136.830.8————
GPT-4oModel Category=Baseline2026.0330.334.424.523.519.837.627.732.831.835.136.830.8————
GPT-4oModel Category=Proprietary LLM2025.1030.3——————————30.82834.3——
GPT-4oWorkflow Mode=Direct Answer, Type=Proprietary2026.0630.3——————————30.828—32.336
ViLaSRModel Category=Agentic, Evaluation Protocol=zero-shot2025.1030.2———————————————
Qwen3-VL-8B-InstructWorkflow Mode=Direct Answer, Type=Open-source2026.0629.8——————————28.130.8—30.127.7
VILASRWorkflow Mode=Direct Answer, Type=Open-source2026.0629.8——————————23.235.9—2621
Bagel-7B-MoTWorkflow Mode=Forced Tool-Use (zero-shot), Type=Open-source2026.0629.7——————————28.731.3—25.624.7
Qwen3-VL-30B-A3B-ThinkingModel Category=Open-weight General Models2026.0629.423.731.935.831.436.522.940.619.718.927.631.3————
InternVL3_5-8BModel Category=Open-weight General Models2026.06292926.629.624.431.825.329.725.814.934.236.4————
VADARModel Category=Agentic, Evaluation Protocol=zero-shot2025.1028.9——————————22.732.826.1——
RoboBrain-2.0Workflow Mode=Direct Answer, Type=Open-source2026.0628.9——————————2828.9—28.822.5
Qwen3-VL-30B-InstructWorkflow Mode=Forced Tool-Use (zero-shot), Type=Open-source2026.0628.9——————————28.131.5—28.721.6
InternVL2.5-8BModel Category=Open-source LLM2025.1028.7——————————30.329.925.4——
Qwen3-VL-8B-InstructWorkflow Mode=Forced Tool-Use (zero-shot), Type=Open-source2026.0628.6——————————30.830.4—29.519.6
Qwen3-VL-8B-ThinkingModel Category=Open-weight General Models2026.0628.631.226.632.129.132.930.15016.717.623.727.3————
InternVL3-78BModel Category=Open-source2026.0328.534.423.432.112.837.626.537.519.728.431.629.3————
InternVL2.5-78BModel Category=Open-source2026.0328.523.722.339.529.131.842.235.919.717.626.327.3————
LLaVA-OneVision-72BModel Category=Open-source2026.0328.44331.933.330.237.638.628.119.713.532.915.7————
LLaVA-3DModel Category=3D-VLM2025.1028——————————19.729.531——
Video-R1Workflow Mode=Direct Answer, Type=Open-source2026.0627.8——————————26.830.5—25.422
VLaserWorkflow Mode=Direct Answer, Type=Open-source2026.0627.3——————————18.929.8—26.926
SpaceRWorkflow Mode=Direct Answer, Type=Open-source2026.0626.9——————————22.529.1—29.421.9
InternVL3-14BModel Category=Open-source LLM2025.1026.8——————————29.325.527.5——
Qwen2.5-VL-3BModel Category=Open-source LLM2025.1026.5——————————23.229.523.2——
Spatial-MLLMWorkflow Mode=Direct Answer, Type=Open-source2026.0626.1——————————26.328.5—25.418
Qwen2.5-VL-7BModel Category=Open-source LLM2025.1025.9——————————25.825.926.1——
SpatialLadderWorkflow Mode=Direct Answer, Type=Open-source2026.0625.4——————————21.230.3—23.316
Video 3D LLMModel Category=3D-VLM2025.1025.3——————————25.825.523.9——
SpatialLLMWorkflow Mode=Direct Answer, Type=Open-source2026.0625.3——————————30.824.5—23.122.7
Llama-3.2-11B-VisionModel Category=Open-source LLM2025.1024.5——————————19.226.923.9——
LLaVA-OneVision-7BModel Category=Open-source LLM2025.1024.5——————————11.62827.1——
Skill-3DModel=GPT-4o2026.06————————————43.2———
Skill-3DModel=GPT-5.42026.06————————————60.4———
Skill-3DModel=Gemini-2.5-Pro2026.06————————————56.7———
Skill-3DModel=Gemini-3-Flash2026.06————————————54.8———
Think3DModel=GPT-4o2026.06————————————38.4———
Think3DModel=GPT-5.42026.06————————————53.4———
Think3DModel=Gemini-2.5-Pro2026.06————————————51———
Think3DModel=Gemini-3-Flash2026.06————————————49.2———
w/ ToolsModel=GPT-4o2026.06————————————34.6———
w/ ToolsModel=GPT-5.42026.06————————————48.1———
w/ ToolsModel=Gemini-2.5-Pro2026.06————————————44.3———
w/ ToolsModel=Gemini-3-Flash2026.06————————————36.8———
w/o ToolsModel=GPT-4o2026.06————————————31.1———
w/o ToolsModel=GPT-5.42026.06————————————42.7———
w/o ToolsModel=Gemini-2.5-Pro2026.06————————————36.9———
w/o ToolsModel=Gemini-3-Flash2026.06————————————32.7———