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

Geometric Question Answering on GEOQA (test)

82.74Total Accuracy

VeriGeo-beam5

Updated 1mo ago

Evaluation Results

Method	Links
VeriGeo-beam5 2026.06		82.74	-	-	-	-
VeriGeo-greedy 2026.06		79.31	-	-	-	-
TR-CoT-Qwen2.5-VL-7B 2026.06		79.2	-	-	-	-
GeoGen-SFT-7B 2026.06		78	-	-	-	-
TR-CoT-InternVL2.5-8B 2026.06		76.7	-	-	-	-
TR-CoT-8B 2026.06		75.9	-	-	-	-
GeoGen-SFT-3B 2026.06		75.6	-	-	-	-
Qwen2.5-VL-7B-Instruct 2026.06		72.07	-	-	-	-
MAVIS-7B 2026.06		68.3	-	-	-	-
G-LLaVA-13B 2026.06		67	-	-	-	-
G-LLaVA-7B 2026.06		62.8	-	-	-	-
GPT-4o 2026.06		61.4	-	-	-	-
InternVL2.5-8B 2026.06		59	-	-	-	-
InternVL2-8B 2026.06		56.5	-	-	-	-
Qwen2-VL-7B/8B 2026.06		55.7	-	-	-	-
FGRPO 2026.06		53.93	53.81	55.4	58.59	50.83
FedProx 2026.06		52.93	56.19	52.17	57.31	50.28
SCAFFOLD 2026.06		51.95	60.48	52.42	54.67	49.75
FedAvg 2026.06		51.4	53.33	45.34	53.79	51.15
FGRPO 2026.06		51.36	60	55.4	59.16	45.34
FedProx 2026.06		48.95	56.67	51.68	52.47	45.84
FedAvg 2026.06		48.87	58.57	50.06	55.33	44.2
SCAFFOLD 2026.06		48.42	57.62	52.17	51.94	44.99
Math-LLaVA-13B 2026.06		47.8	-	-	-	-
GPT-4V 2026.06		43.4	-	-	-	-
FGRPO 2026.06		40.76	47.62	45.34	46.65	35.88
FedProx 2026.06		39.28	47.62	40.99	44.89	35.09
SCAFFOLD 2026.06		38.08	51.9	41.37	40.79	34.98
FedAvg 2026.06		35.59	41.43	36.89	40.93	31.78
DeepSeek-VL-7B 2026.06		33.7	-	-	-	-
FGRPO 2026.06		28.42	36.19	31.18	30.13	26.37
FedAvg 2026.06		26.82	33.81	30.06	27.49	25.34
SCAFFOLD 2026.06		26.75	35.24	29.81	29.38	24.04
FedProx 2026.06		26.41	32.86	30.68	27.49	24.49