GUI Grounding on ScreenSpot v2
96.5Avg AccuracyMAI-UI-32B
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MAI-UI-32Bparameters=32B2025.12 | 96.5 | 99 | 92.9 | 99.5 | 93.6 | 97.4 | 94.6 | — | — | — | — | — | |
| MAI-UI-32BParameters=32B2026.06 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | |
| OmegaUse-GBackbone=30B-A3B VL, Training Protocol=RL2026.01 | 96.3 | 99.3 | 94.3 | 99 | 96.4 | 97.4 | 94 | — | — | — | — | — | |
| UI-Venus-1.5-30B-A3BVersion=1.5, Parameters=30B-A3B2026.06 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-1.5-8BAccess=Open weights2026.03 | 95.9 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-1.5-8BVersion=1.5, Parameters=8B2026.06 | 95.9 | — | — | — | — | — | — | — | — | — | — | — | |
| POINTS-GUI-G-8BParameters=8B2026.02 | 95.7 | — | — | — | — | — | — | — | — | — | — | — | |
| quality-aware self-distillation2026.06 | 95.68 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-Ground-72BModel Category=GUI-specific, Training Protocol=RL2026.01 | 95.3 | 99.7 | 93.8 | 95.9 | 90 | 96.2 | 92.6 | — | — | — | — | — | |
| UI-Venus-72BParameters=72B2026.02 | 95.3 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-72Bparameters=72B2025.12 | 95.3 | 99.7 | 93.8 | 95.9 | 90 | 96.2 | 92.6 | — | — | — | — | — | |
| UI-Venus-Ground-72BParameters=72B2026.06 | 95.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-1.5-32B-InstructVersion=1.5, Parameters=32B, Variant=Instruct2026.06 | 95.3 | — | — | — | — | — | — | — | — | — | — | — | |
| Seed 1.5-VL2025.05 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| Seed1.5-VLModel Category=Closed-source2026.01 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| MAI-UI-8BParameters=8B2026.02 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GTA1-32BParameters=32B2026.02 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GTA1-32Bparameters=32B2025.12 | 95.2 | 99.7 | 90.5 | 99 | 94.3 | 95.7 | 90.1 | — | — | — | — | — | |
| MAI-UI-8Bparameters=8B2025.12 | 95.2 | 99.3 | 89.1 | 99 | 92.1 | 97.9 | 91.1 | — | — | — | — | — | |
| SeedVL-1.5Version=1.52025.12 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| MAI-UI-8BAccess=Open weights2026.03 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| Seed-1.5-VLModel Category=Closed-Source Models2026.04 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| ZoomOnceSize=4B2026.06 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| MAI-UI-8BParameters=8B2026.06 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | |
| Step-GUI-8BParameters=8B2025.12 | 95.1 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-SDTime/epoch=4.22026.05 | 95.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Step-GUI-8BParameters=8B2026.06 | 95.1 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-SD2026.06 | 94.97 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-InsSize=7B2026.06 | 94.9 | — | — | — | — | — | — | — | — | — | — | — | |
| GTA1-72BModel Category=GUI-specific, Training Protocol=RL2026.01 | 94.8 | 99.3 | 92.4 | 97.4 | 89.3 | 95.3 | 91.4 | — | — | — | — | — | |
| MAI-UISize=4B2026.06 | 94.8 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-30B-A3B-InstructParameters=30B, Arch=A3B, Type=Instruct2025.12 | 94.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Xiaomi-GUI-0-30B-A3BParameters=30B-A3B2026.06 | 94.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-Instruct + GRPO-BinaryTime/epoch=16.92026.05 | 94.6 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-32BParameters=32B2026.02 | 94.5 | 98.28 | 89.1 | 98.97 | 90.71 | 94.87 | 92.61 | — | — | — | — | — | |
| Qwen-32BBackbone=Qwen3-VL-Thinking-32B2026.03 | 94.5 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-8B-InstructParameters=8B, Type=Instruct2025.12 | 94.4 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-TARS 1.52025.05 | 94.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-TARS-1.5Model Category=Closed-source2026.01 | 94.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-TARS-1.5Version=1.52025.12 | 94.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-TARS 1.5-7BAccess=Open weights2026.03 | 94.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-Ground-7BModel Category=GUI-specific, Training Protocol=RL2026.01 | 94.1 | 99 | 90 | 97 | 90.7 | 96.2 | 88.7 | — | — | — | — | — | |
| UI-Venus-7BParameters=7B2026.02 | 94.1 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-7Bparameters=7B2025.12 | 94.1 | 99 | 90 | 96.9 | 90.7 | 96.2 | 88.7 | — | — | — | — | — | |
| UI-VenusSize=7B2026.06 | 94.1 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Venus-Ground-7BParameters=7B2026.06 | 94.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-4B-InstructParameters=4B, Type=Instruct2025.12 | 94 | — | — | — | — | — | — | — | — | — | — | — | |
| UI-Ins-7B2026.05 | 94 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-Instruct + GRPO-GaussianTime/epoch=16.82026.05 | 94 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-32B-InstructContext Window=32K2026.03 | 93.9 | — | — | — | — | — | — | — | — | — | — | — | |
| MolmoPoint-GUI-8BGUI-specific Fine-tuning=true, Test-time scaling=64crops2026.03 | 93.9 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-CursorModel Category=GUI-specific Models2026.04 | 93.9 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Cursor-7BTraining Paradigm=Reinforcement Learning, Parameter Scale=7B, Backbone=Qwen2.5-VL-7B2025.09 | 93.9 | — | — | — | — | — | — | — | 94.9 | 92.9 | 92.6 | — | |
| GUI-Cursor-7BTraining Paradigm=Reinforcement Learning, Parameter Scale=7B, Backbone=UI-TARS-1.5-7B2025.09 | 93.9 | — | — | — | — | — | — | — | 93.3 | 95.8 | 91.7 | — | |
| Naive-OPSD2026.06 | 93.87 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-Instruct + GRPO-DistanceTime/epoch=16.72026.05 | 93.8 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-32B-InstructContext Window=4K2026.03 | 93.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Gemini-3-ProAccess=API call only2026.03 | 93.7 | — | — | — | — | — | — | — | — | — | — | — | |
| HyperClick-7B2026.05 | 93.7 | — | — | — | — | — | — | — | — | — | — | — | |
| HyperClickSize=7B2025.10 | 93.7 | 98.3 | 93.4 | 96.9 | 85.7 | 96.2 | 86.7 | — | — | — | — | — | |
| GUI-Owl-1.5-8B-InstructVersion=1.5, Parameters=8B, Variant=Instruct2026.06 | 93.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL ThinkingNumber of Parameters=8B2026.01 | 93.6 | — | — | — | — | — | — | — | — | — | — | — | |
| Step-GUI-4BParameters=4B2025.12 | 93.6 | — | — | — | — | — | — | — | — | — | — | — | |
| Step-GUI-4BParameters=4B2026.06 | 93.6 | — | — | — | — | — | — | — | — | — | — | — | |
| InfiGUI-G1-7BParameters=7B2026.02 | 93.5 | — | — | — | — | — | — | — | — | — | — | — | |
| InfiGUI-G1-7Bevaluated_by_authors=true, parameters=7B2025.12 | 93.5 | 99 | 91.9 | 94.3 | 82.1 | 97.9 | 89.2 | — | — | — | — | — | |
| InfiGUI-G1Size=7B2026.06 | 93.5 | — | — | — | — | — | — | — | — | — | — | — | |
| EvoCUA-OpenCUA-72BBase Model=OpenCUA-72B2026.01 | 93.47 | — | — | — | — | — | — | — | — | — | — | — | |
| OpenCUA-32BModel Category=GUI-specific, Training Protocol=SFT2026.01 | 93.4 | — | — | — | — | — | — | — | — | — | — | — | |
| OpenCUA-32BParameters=32B2026.02 | 93.4 | — | — | — | — | — | — | — | — | — | — | — | |
| OpenCUA-32Bparameters=32B2025.12 | 93.4 | — | — | — | — | — | — | — | — | — | — | — | |
| MolmoPoint-GUI-8BGUI-specific Fine-tuning=true2026.03 | 93.4 | — | — | — | — | — | — | — | — | — | — | — | |
| OpenCUASize=32B2026.06 | 93.4 | — | — | — | — | — | — | — | — | — | — | — | |
| SFT2026.06 | 93.32 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2Training Data Size=100K, Model Size=7B, Training Paradigm=RL2026.02 | 93.3 | 98.3 | 91.9 | 95.4 | 89.3 | 94 | 87.7 | — | — | — | — | — | |
| GUI-G2-7BModel Category=GUI-specific, Training Protocol=RL2026.01 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2-7BParameters=7B2026.02 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2-7Bevaluated_by_authors=true, parameters=7B2025.12 | 93.3 | 98.3 | 91.9 | 95.4 | 89.3 | 94 | 87.7 | — | — | — | — | — | |
| GUI-G²Evaluation Protocol=Upper Bound2026.01 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2-VL-7B-ThinkingMax Output Tokens=4K2026.03 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2-7BModel Category=GUI-specific Models2026.04 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2-7B2026.05 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G^2-7BTraining Paradigm=Reinforcement Learning, Parameter Scale=7B2025.09 | 93.3 | — | — | — | — | — | — | — | 95.1 | 92.4 | 90.9 | — | |
| GUI-G2Size=7B2025.10 | 93.3 | 98.3 | 91.9 | 95.4 | 89.3 | 94 | 87.7 | — | — | — | — | — | |
| Upper BoundNote=Reference Line 22026.06 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-G2Size=7B2026.06 | 93.3 | — | — | — | — | — | — | — | — | — | — | — | |
| ReCAP-32BBackbone=Qwen3-VL-Thinking-32B, Status=Finetuned2026.03 | 93.24 | — | — | — | — | — | — | — | — | — | — | — | |
| Trifuse+GUI-ActorModel Size=7B, Training Paradigm=Attention-based Training-Based2026.02 | 93.2 | 98.2 | 90.8 | 97.4 | 87.1 | 93.9 | 88.3 | — | — | — | — | — | |
| GUI-Owl-32BParameters=32B2026.02 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-32Bparameters=32B2025.12 | 93.2 | 98.6 | 90 | 97.9 | 87.8 | 94.4 | 86.7 | — | — | — | — | — | |
| GUI-Owl-32BParameters=32B2025.12 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| Holo2-8BAccess=Open weights2026.03 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-Instruct2026.05 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-OwlSize=32B2026.06 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-32BParameters=32B2026.06 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-1.5-4B-InstructVersion=1.5, Parameters=4B, Variant=Instruct2026.06 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-1.5-8B-ThinkingVersion=1.5, Parameters=8B, Variant=Thinking2026.06 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| GUI-Owl-1.5-32B-ThinkingVersion=1.5, Parameters=32B, Variant=Thinking2026.06 | 93.2 | — | — | — | — | — | — | — | — | — | — | — | |
| FOCUSUI-7BParameters=7B, Visual token retention ratio (r)=100%2026.01 | 93.1 | 98.8 | 91.6 | 95.6 | 92.1 | 95 | 84.4 | — | — | — | — | — | |
| Qwen3-VL-32BParameters=32B2026.02 | 93 | — | — | — | — | — | — | — | — | — | — | — |