Visual Reasoning on CharXiv (val) benchmark leaderboard · SOTA2 Research

Visual Reasoning on CharXiv (val)

37.95Text in Chart AccuracyVReST-Vote

Evaluation Results

Method	Links
VReST-Vote2025.06		37.95	61.62	39.22	27.07	37.3	38.41	45.38	43.7	38.58	31.75	36.21	32.74	38.1
VReST2025.06		34.32	51.52	36.21	26.2	33.33	36.96	35.29	36.3	36.22	33.33	31.03	33.63	34.6
VReST-Vote2025.06		34.32	52.53	35.78	31	34.92	39.86	36.97	34.07	40.16	35.71	30.17	32.74	35.7
Best-of-N2025.06		34.09	48.48	28.02	24.02	33.33	30.43	30.25	35.56	38.58	31.75	24.14	29.2	31.8
ToT2025.06		34.09	45.45	33.62	20.96	30.95	26.81	36.97	31.85	35.43	29.37	26.72	39.82	32.1
VReST2025.06		33.64	54.55	33.62	22.27	30.95	31.16	41.18	40.74	33.86	26.98	30.17	29.2	33.1
ToT2025.06		33.41	47.47	37.93	24.45	33.33	36.23	30.25	34.81	37.01	33.33	31.03	33.63	33.8
CoT-Vote2025.06		32.95	45.45	28.88	22.71	26.98	28.99	33.61	30.37	39.37	29.37	25.86	32.74	30.9
Best-of-N2025.06		32.73	48.48	34.05	25.33	31.75	34.06	29.41	34.07	35.43	33.33	33.62	30.97	32.9
QA2025.06		31.82	38.38	28.45	22.27	33.33	30.43	31.93	29.63	35.43	25.4	21.55	27.43	29.5
CoT-Vote2025.06		31.82	49.49	33.19	24.02	33.33	34.06	30.25	31.85	36.22	32.54	27.59	30.09	32.1
QA2025.06		30.45	47.47	31.9	21.83	30.95	32.61	27.73	31.11	34.65	30.95	27.59	27.43	30.5
COT2025.06		29.09	40.4	26.72	18.78	21.43	27.54	32.77	29.63	26.77	23.81	23.28	33.63	27.3
CoT2025.06		28.18	44.44	29.31	20.96	28.57	29.71	25.21	28.89	33.86	27.78	26.72	25.66	28.4
Cantor2025.06		28.18	44.44	31.47	21.4	29.37	31.88	26.05	30.37	33.07	28.57	25	26.55	29
Cantor2025.06		27.73	43.43	30.6	23.58	26.19	27.54	27.73	31.11	37.01	24.6	30.17	27.43	29