Multimodal Intent Recognition on MIntRec benchmark leaderboard · SOTA2 Research

Multimodal Intent Recognition on MIntRec

86.29AccuracyQwen-VL-72B

Evaluation Results

Method	Links
Qwen-VL-72B2026.03		86.29	—	—	86.09	—	86.75	—	—
Qwen-2VL-7B2026.03		82.92	—	—	82.79	—	86.75	—	—
MiniCPM-V-2.6-8B2026.03		80.67	—	—	80.56	—	82.19	—	—
LLaVA-Video-72B2026.03		80.22	—	—	79.94	—	80.63	—	—
QA-MoE2026.04		75.3	—	72.2	—	—	—	—	—
Atsuko2026.03		74.16	—	71.22	—	72.93	—	70.86	—
GCL2026.05		72.74	—	70.95	—	72.31	—	71.24	—
MulT2026.04		72.6	—	69.5	—	—	—	—	—
EMOE2026.04		72.6	—	70.7	—	—	—	—	—
GsiT2026.05		72.6	—	69.4	—	69.4	—	70.1	—
TSDA2026.05		72.59	—	70.68	—	71.97	—	70.75	—
MulT2026.03		72.58	—	69.36	—	70.73	—	69.47	—
EMOE2026.03		72.58	—	70.73	—	72.08	—	70.86	—
MuLT2026.05		72.58	—	69.36	—	70.73	—	69.47	—
EMOE2026.05		72.58	—	70.73	—	72.08	—	70.86	—
CAGC2026.05		72.53	—	70.62	—	70.86	—	70.55	—
MISA2026.04		72.4	—	70.8	—	—	—	—	—
MISA2026.03		72.36	—	70.57	—	71.24	—	70.41	—
MISA2026.05		72.36	—	70.57	—	71.24	—	70.41	—
MAG-BERT2026.03		70.34	—	68.19	—	68.31	—	69.36	—
MAG-BERT2026.05		70.34	—	68.19	—	68.31	—	69.36	—
GPT-42026.03		63.84	—	—	63.97	—	68.16	—	—
Nano-EmoX2026.03		58.2	—	—	58.17	—	60.12	—	—
Nano-EmoX‡2026.03		56.18	—	—	56.62	—	65.05	—	—
AffectGPT (s)2026.03		48.76	—	—	50.71	—	57.87	—	—
MobileVLM V2-3B2026.03		36.12	—	—	37.74	—	35.12	—	—
MobileVLM V2-1.7B2026.03		33.27	—	—	33.81	—	34.75	—	—
GsiT2025.05		—	72.6	69.4	72.7	69.4	73.5	70.1	72.6
MMIM2025.05		—	70.8	68.7	71	69.2	71.8	68.9	70.8
MulT2025.05		—	71.2	68.2	71.1	68.9	71.4	68.1	71.2