Motion-to-Text on HumanML3D benchmark leaderboard · SOTA2 Research

Motion-to-Text on HumanML3D

60.2R@1MotionMERGE

Evaluation Results

Method	Links
MotionMERGE2026.05		60.2	88	2.51	—	9.44	—	—	38.8	—	—
MG-MotionLLM2026.03		59.2	86.6	2.581	—	8.06	—	—	36.7	—	—
MG-MotionLLM2026.05		59.2	86.6	2.581	—	8.06	—	—	36.7	—	—
UniMotion2026.03		56.2	85.7	2.481	60.4	20.7	46.5	39.3	41.2	—	—
MotionGPT-22024.10		55.8	83.8	2.767	48.7	13.8	37.6	29.8	32.6	73.8	15.27
LaMPM2T2026.03		54.7	83.1	2.808	47.8	13.04	37.1	28.9	32.7	—	—
UniMotion†2026.03		54.7	84.3	2.562	56.9	18.6	44.7	35.4	38.5	—	—
LaMP-M2T2026.05		54.7	83.1	2.808	—	13.04	—	—	32.7	—	—
MotionGPT2026.03		54.3	82.7	2.821	48.2	12.47	37.4	29.2	32.4	—	—
MotionGPT2026.05		54.3	82.7	2.821	—	12.47	—	—	32.4	—	—
MotionGPT2024.10		54.3	82.7	2.821	48.2	12.5	37.4	29.2	32.4	—	13.04
Real2026.03		52.3	82.8	2.901	—	—	—	—	—	—	—
Real2026.05		52.3	82.8	2.901	—	—	—	—	—	—	—
Real Desc2024.10		52.3	82.8	2.901	—	—	—	—	—	72.5	—
TM2T2026.03		51.6	82.3	2.935	48.9	7	38.1	16.8	32.2	—	—
TM2T2026.05		51.6	82.3	2.935	—	7	—	—	32.2	—	—
TM2T2024.10		51.6	82.3	2.935	48.9	7	38.1	16.8	32.2	72	10.67
Seq2Seq(Att)2024.10		43.6	70.6	3.447	51.8	17.9	46.4	58.4	29.1	61.1	—
SeqGAN2024.10		33.2	53.2	4.895	47.8	13.5	39.2	50.2	23.4	45.7	—
RAEs2024.10		10	26.1	6.337	33.3	10.2	37.5	22.1	10.7	18.8	—
MG-MotionLLM2026.03		—	—	—	—	—	—	—	36.7	—	—
MotionGPT2026.03		—	—	—	—	—	—	—	32.4	—	—
UniMotion2026.03		—	—	—	—	—	—	—	41.2	—	—