Multimodal Perception and Reasoning on FCMBench 1.0 (test) benchmark leaderboard · SOTA2 Research

Multimodal Perception and Reasoning on FCMBench 1.0 (test)

94.22DTRQfin-VL-Instruct

Evaluation Results

Method	Links
Qfin-VL-Instruct2026.01		94.22	45.38	55	79.32	76.73	50.64	53.14	64.92
Qwen3-VL-235B-A22B-Instruct2026.01		92.35	38.58	46.2	70.81	59.33	43.94	49.67	57.27
Gemini 3 Pro2026.01		90.66	35.46	48.8	66	89.11	52.36	69.9	64.61
Claude Opus 4.52026.01		87.25	30.93	41.53	73.24	83.59	40.77	62.06	59.91
Qwen3-VL-32B-Instruct2026.01		78.48	33.28	46.4	77.2	70.95	37.82	53.66	56.83
Ovis2.52026.01		77.5	24.13	41.4	66.02	67.02	31.04	59.06	52.31
InternVL-3.5-38B2026.01		77.21	22.63	36.4	40.86	58.24	27.6	45.06	44
GLM-4.5V2026.01		75.1	29.61	35	68.55	74.45	43.55	44.29	52.94
GPT 5.12026.01		73.94	21.38	39.6	47.4	67.61	33.6	45.32	46.98
InternVL-3.5-241B-A28B2026.01		72.76	23.77	34.2	53.63	65.82	36.15	46.56	47.56
GPT 5.22026.01		72.19	25.7	43.8	70.82	70.42	39.44	49.6	53.14
MiniCPM-V 4.52026.01		69.01	12.55	31.6	47.38	55.75	21.93	46.56	40.68
Qwen3-VL-8B-Instruct2026.01		67.69	33.12	40	61.94	52.4	35.63	51.76	48.93
InternVL-3.5-8B2026.01		61.53	22.08	28.8	55.45	68.64	23	46.56	43.72
InternVL-3.5-30B-A3B2026.01		61.36	21.21	34.8	50.83	66.42	24.14	49.67	44.06
Qwen3-VL-30B-A3B-Instruct2026.01		60.95	28.17	30.2	45.32	62.76	25.16	51.29	43.41
Llama-4-Maverick2026.01		54.98	20.89	41.8	38.71	69.1	30.46	42.27	42.6
MiniMax-VL-012026.01		54.57	21.11	44.6	47.28	70.23	36.6	46.98	45.91
Kimi-VL2026.01		51.09	26.29	26.2	50.42	61.21	20.79	46.56	40.37
Grok 42026.01		50.85	15.41	36	41.36	65.42	25.09	45.71	39.98
DeepSeek-OCR + DeepSeek V3.22026.01		48.61	20.97	12.2	31.98	45.67	45.47	51.21	36.59
LLaVA-OneVision-1.52026.01		39.53	21.14	27.6	38.62	50.94	44.17	46.56	38.37
Phi-4-multimodal-instruct2026.01		17.36	13.59	9.2	38.85	37.66	24.73	47.48	26.98
Janus-Pro2026.01		8.73	7.2	10.4	59.22	44.97	21.48	46.56	28.37