Text-to-Video Generation on VBench (Total Score and Detailed Semantic Metrics) benchmark leaderboard · SOTA2 Research

Text-to-Video Generation on VBench (Total Score and Detailed Semantic Metrics)

82.65Total ScoreRAPO++

Evaluation Results

Method	Links
RAPO++2025.10		82.65	97.46	73.48	99.2	98.78	71.89	64.76
RAPO2025.10		82.38	96.86	71.4	96.8	96.91	64.86	59.15
LaVie2025.10		80.89	96.62	69	95.8	92.09	37.71	37.27
RAPO++2025.10		80.75	97.93	67.84	97.2	93.82	55.38	46.87
RAPO2025.10		79.97	98.17	66.72	95.2	96.47	52.78	41.31
Prompt Refiner2025.10		79.75	96.42	70.42	87	91.29	36.52	54.37
GPT42025.10		79.69	96.14	70.27	83.8	88.73	36.23	50.55
PAE2025.10		79.17	96.58	70.32	82.4	73.23	42.54	37.92
Promptist2025.10		79.13	96.63	70.08	81	71.04	43.97	37.76
GPT42025.10		77.4	97.52	63.54	85.8	78.32	27.73	36.72
Prompt Refiner2025.10		77.23	97.67	64.19	84.6	83.6	30	35.12
Latte2025.10		77.03	97.1	63.38	88.4	83.86	29.55	40.63
PAE2025.10		76.83	97.78	63.52	76.4	81.52	23.42	33.49
Promptist2025.10		76.65	97.82	63.37	74	79.18	21.72	31.31