Share your thoughts, 1 month free Claude Pro on usSee more

GUI Grounding on ScreenSpot Mobile V2

100Text Accuracy

UI-AGILE-7B

Updated 3d ago

Evaluation Results

Method
UI-AGILE-7B 2025.07	100	91.1	-
GUI-RC 2026.04	99.9	85.9	-
GTA1-32B 2026.02	99.7	90.5	-
UI-Venus-72B 2026.02	99.7	93.8	-
GTA1-7B 2026.04	99.7	90.5	-
UI-R1-E 2025.07	99.6	80.1	-
UI-AGILE-3B 2025.07	99.6	86.4	-
SE-GUI-7B 2026.02	99.3	89.1	-
MAI-UI-2B 2026.02	99.3	87.2	-
MAI-UI-8B 2026.02	99.3	89.1	-
GUI-G2-7B 2026.04	99.3	91.5	96
GTA1-7B 2026.02	99	88.6	-
GUI-Owl-7B 2026.02	99	92.4	-
InfiGUI-G1-7B 2026.02	99	91.9	-
UI-Venus-7B 2026.02	99	90	-
POINTS-GUI-G-8B 2026.02	99	91	-
GTA1-3B 2026.04	99	88.6	-
UI-Venus-7B 2026.04	99	90.1	95.2
GUI-R1-7B 2025.07	98.8	86.4	-
GUI-Owl-32B 2026.02	98.6	90	-
UI-Venus-7B 2026.04	98.6	90.5	95.2
GUI-G2-7B 2026.04	98.6	91.5	95.6
Qwen2.5-VL-7B 2025.07	98.4	84.8	-
GUI-G2-7B 2026.02	98.3	91.9	-
Qwen2.5-VL-32B 2026.04	98.3	86.7	-
UI-R1-E-3B 2026.04	98.2	83.9	-
GUI-R1-3B 2025.07	98.1	79	-
Qwen3-VL-8B 2026.02	97.9	84.8	-
Qwen2.5-VL-7B 2026.04	97.9	87.2	93.4
GUI-Actor-7B 2026.02	97.6	88.2	-
Qwen2.5-VL-7B 2026.04	97.2	86.7	92.8
UI-TARS-7B 2026.02	96.9	89.1	-
UI-TARS-7B 2025.07	96.9	89.1	-
UI-TARS-7B 2026.04	96.9	89.1	-
Jedi-7B 2026.04	96.9	87.2	-
Jedi-3B 2026.04	96.6	81.5	-
UI-TARS-1.5-7B 2026.04	96.5	86.4	-
Qwen3-VL-32B 2026.02	96.2	90	-
Qwen2.5-VL-3B 2025.07	96.1	74.8	-
Qwen3-VL-2B 2026.02	95.5	82	-
OS-Atlas-7B 2026.02	95.2	75.8	-
UI-TARS-2B 2025.07	95.2	79.1	-
OS-Atlas-7B 2026.04	95.2	75.8	78.3
OS-Atlas-7B 2025.07	95	73.3	-
Aguvis-7B 2025.07	94.9	80.1	-
UI-TARS-72B 2025.07	94.8	86.3	-
UI-TARS-72B 2026.04	94.8	86.3	-
DiMo-GUI 2026.04	94.8	85.3	-
Qwen2.5-VL-3B 2026.04	93.4	73.5	-
UI-Tars-1.5-7B 2026.02	92.2	81.5	-
Phi-ground 2026.02	90.2	76.4	-
OS-Atlas-4B 2025.07	87.2	59.7	-
OS-Atlas-4B 2026.04	87.2	59.7	74.9
UGround-v1-7B 2026.02	83.6	90.5	-
SeeClick 2025.07	78.4	50.7	-
GPT-4o 2026.04	-	-	22.5
UGround-7B 2026.04	-	-	74.3
SE-GUI-7B 2026.04	-	-	95.2