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

GUI Grounding on Screenspot v1

91.67Average Accuracy

Qwen3-VL-32B

Updated 20d ago

Evaluation Results

Method
Qwen3-VL-32B 2026.02	91.67	96.34	85.59	96.91	87.86	90.87	90.78
GUI-Actor-7B (Qwen2.5-VL) 2025.11	89.9	96.3	85.2	95.4	82.9	90.4	85.4
UI-TARS-7B 2025.11	89.5	94.5	85.2	95.9	85.7	90	83.5
UGround-V1-72B 2025.11	89.4	94.1	83.4	94.9	85.7	90.4	87.9
MFT 2026.02	89.2	93.4	82.8	96.4	84.3	90.8	89.4
Aguvis-72B 2025.11	89.2	94.5	85.2	95.4	77.9	91.3	85.9
UI-TARS-72B 2026.02	88.92	94.14	82.53	95.36	85	90.87	83.5
GUI-AIMA-3B-lite 2025.11	88.8	97.1	82.1	95.4	83.6	90.9	80.6
Ovis2.5S-RF++ 2026.02	88.68	95.24	79.04	94.85	84.29	92.61	83.5
Ovis2.5SFT 2026.02	88.6	95.24	78.6	94.85	85	92.61	83.01
Ovis2.5S-GRPO 2026.02	88.52	94.51	79.04	94.85	85	92.61	83.01
Qwen3-VL-8B 2026.02	88.44	95.24	79.91	94.85	81.43	92.17	83.5
Ovis2.5S-RLOO 2026.02	88.44	94.87	79.04	94.85	85	92.17	82.52
UI-TARS-72B 2026.02	88.4	94.9	82.5	89.7	88.6	88.7	85
GUI-Actor-3B (Qwen2.5-VL) 2025.11	88.4	94.5	83.8	92.8	82.1	91.3	82.5
UI-TARS-72B 2025.11	88.4	94.9	82.5	89.7	88.6	88.7	85
GUI-Actor-7B 2026.02	88.3	94.9	82.1	91.8	80	91.3	85.4
GUI-Actor-7B (Qwen2-VL) 2025.11	88.3	94.9	82.1	91.8	80	91.3	85.4
UI-TARS-7B 2026.02	88.21	92.31	85.15	90.72	84.29	88.7	85.92
SparkUI-Parser 2025.11	88	94.9	83.8	95.9	80.7	89.6	82.9
GUI-Actor-2B (Qwen2-VL) 2025.11	86.5	93	79.9	88.1	78.6	90.9	84
UGround-v1-7B 2025.11	86.3	93	79.9	93.8	76.4	90.9	84
Claude 4 Sonnet CU 2026.02	84.75	97.07	79.48	81.44	64.29	92.61	82.52
Aguvis-7B 2025.11	84.4	95.6	77.7	93.8	67.1	88.3	75.2
UI-TARS 1.5-7B 2026.02	84.2	92.67	77.73	93.81	71.43	86.52	77.18
Gemini 2.0 2025.11	84	-	-	-	-	-	-
Claude Computer Use 2025.11	83	-	-	-	-	-	-
OS-Atlas-7B 2025.11	82.5	93	72.9	91.8	62.9	90.9	74.3
UI-TARS-2B 2025.11	82.3	93	75.5	90.7	68.6	84.3	74.8
Aguvis-G-7B 2025.11	81.8	88.3	78.2	88.1	70.7	85.7	74.8
Claude 3.7 Sonnet CU 2026.02	80.74	88.28	73.36	79.9	62.86	90.87	80.58
UGround-V1-2B 2025.11	77.7	89.4	72	88.7	65.7	81.3	68.9
Magma-8B 2025.11	60.3	60.4	58.5	75.3	52.9	69.1	52
Qwen2-VL-7B 2025.11	55.3	75.5	60.7	76.3	54.3	35.2	25.7
SeeClick-9.6B 2025.11	53.4	78	52	72.2	30	55.7	32.5
CogAgent-7B 2025.11	47.4	67	24	74.2	20	70.4	28.6
GPT-4o 2026.02	18.3	20.2	24.9	21.1	23.6	12.2	7.8
GPT-4o 2025.11	18.3	20.2	24.9	21.1	23.6	12.2	7.8
GPT-4 2025.11	16.2	22.6	24.5	20.2	11.8	9.2	8.8