Multi-modal Visualization Generation on MultiVis-Bench Code-Referenced Generation benchmark leaderboard · SOTA2 Research

Multi-modal Visualization Generation on MultiVis-Bench Code-Referenced Generation

78.53Low-Level Type ScoreMultiVis-Agent

Evaluation Results

Method	Links
MultiVis-Agent2026.01		78.53	44.1	41.59	86.65	92.26	77.18	70.05	89.65	89.03	79.06	73.55	82.79	95.44	83.11	76.58
MultiVis-Agent2026.01		70.13	35.45	20.62	79.96	78.61	73.22	59.67	79.87	80.37	70.13	55.68	65.22	89.94	70.9	65.28
LLM Workflow2026.01		66.67	40.85	26.2	63.58	65.13	68.59	55.17	77.89	78.34	66.38	61.8	68.98	81.91	70.72	62.94
Instructing LLM2026.01		65.32	45.47	32.37	76.88	75.92	68.4	60.73	76.59	76.71	65.32	64.21	70.13	82.1	70.71	65.72
LLM Workflow2026.01		64.74	36.8	12.91	63.2	39.31	65.32	47.05	71.34	70.71	61.95	51.69	58.29	76.53	63.09	55.07
Instructing LLM2026.01		63.39	39.31	16.57	73.22	47.78	63.58	50.64	72.3	73.99	63.87	57.85	62.91	79.08	66.52	58.58
nvAgent2026.01		26.78	15.41	2.31	56.65	61.66	50.29	35.52	30.64	29.34	30.25	19.99	32.66	52.04	30.73	33.12
nvAgent2026.01		22.74	19.27	2.31	50.1	53.18	40.27	31.31	29.38	28.71	27.46	21.19	31.84	45.13	29.26	30.29