🖼️ Multimodal League

#	Model	AIM	Năng lực
1	Gemini 3.5 Flash Google	94.7	100	NEW
2	Qwen3.6 Plus Alibaba	94.7	100	NEW
3	GPT-5.1 OpenAI	93.0	98	NEW
4	o3 OpenAI	90.3	95	NEW
5	GPT-5 OpenAI	89.6	94	NEW
6	Qwen3.5-122B-A10B Alibaba Cloud/Qwen Team	88.7	93	NEW
7	Qwen3.6-27B Alibaba Cloud/Qwen Team	85.9	90	NEW
8	o4-mini OpenAI	85.4	89	NEW
9	Qwen3.5-27B Alibaba Cloud/Qwen Team	84.2	88	NEW
10	Gemini 2.5 Pro Google	83.3	87	NEW
11	Qwen3.6-35B-A3B Alibaba Cloud/Qwen Team	82.5	86	NEW
12	Qwen3.5-35B-A3B Alibaba Cloud/Qwen Team	81.6	85	NEW
13	Gemini 2.5 Flash Google	76.8	79	NEW
14	o1 OpenAI	74.5	77	NEW
15	Claude Sonnet 4 Anthropic	61.8	62	NEW
16	GPT-4.1 OpenAI	60.9	61	NEW
17	Llama 4 Maverick Meta	60.7	61	NEW
18	GPT-4.1 mini OpenAI	59.0	59	NEW
19	Gemini 2.5 Flash-Lite Google	57.6	57	NEW
20	Llama 4 Scout Meta	51.4	50	NEW
21	Qwen2.5 VL 72B Instruct Alibaba Cloud/Qwen Team	49.9	48	NEW
22	Qwen2.5 VL 32B Instruct Alibaba Cloud/Qwen Team	49.4	48	NEW
23	Claude 3.5 Sonnet Anthropic	44.6	42	NEW
24	Mistral Small 3.2 24B Mistral AI	37.2	34	NEW
25	DeepSeek VL2 DeepSeek	35.7	32	NEW
26	GPT-4o OpenAI	33.9	30	NEW
27	DeepSeek VL2 Small DeepSeek	30.6	26	NEW
28	Nova Pro Amazon	25.8	21	NEW
29	gpt-4o-mini OpenAI	20.0	14	NEW
30	Mistral Small 3.1 24B Mistral AI	19.0	13	NEW
31	Qwen2.5 VL 7B Instruct Alibaba Cloud/Qwen Team	17.1	11	NEW
32	GPT-4.1 nano OpenAI	13.8	7	NEW
33	DeepSeek VL2 Tiny DeepSeek	13.2	6	NEW
34	Nova Lite Amazon	10.3	3	NEW
35	Gemini 3.1 Pro Preview Google	8.0	0	NEW
36	Llama 3.2 11B Meta	8.0	0	NEW

● Năng lực 65%● Phổ biến (mạng xã hội) 25%● Độ mới 10%

MMMUMathVista

Nguồn dữ liệu thật (3)

Điểm minh hoạ (seed). Pipeline R3/R4 sẽ thay bằng AIM Index thật từ benchmark + mạng xã hội.