AI Model Leaderboard

Compare 107+ large language models side-by-side across five key benchmarks. Find the best model for your use case — sorted by overall performance, cost, or capability.

107

Total Models

Frontier Models

Open Source

#	Model	Provider	MMLU	HumanEval	Math	Reasoning	Coding	Overall ↓	Context
1	OpenAI o3 Multimodal	OpenAI	93.8	94.5	96.2	96.8	94	95.1	200K
2	GPT-5.2 Multimodal	OpenAI	95.2	95.8	93.1	95.5	95.3	95	512K
3	Claude Opus 4.6 Multimodal	Anthropic	94.8	95.5	91.8	95	95.2	94.5	400K
4	Gemini 3 Ultra Multimodal	Google	94.5	93	92.5	94.8	93.2	93.6	2M
5	DeepSeek R2 Open Source	DeepSeek	92	93.5	95	94.5	93	93.6	256K
6	OpenAI o1	OpenAI	91.8	92.4	94.8	95.1	92	93.2	200K
7	GPT-5 Multimodal	OpenAI	93.5	94	89.5	94.2	93.8	93	256K
8	DeepSeek R1 Open Source	DeepSeek	90.8	92	94.3	93.8	91.5	92.5	128K
9	Grok 4 Multimodal	xAI	93	92.5	91	93.5	92	92.4	256K
10	Claude Opus 4 Multimodal	Anthropic	92.3	95.2	85.6	93.5	95	92.3	200K
11	Gemini 2.5 Pro Multimodal	Google	92	93.2	86.4	92.8	93	91.5	1M
12	OpenAI o4-mini	OpenAI	88.2	91.5	92.8	92	91	91.1	200K
13	Llama 4 Behemoth Open SourceMultimodal	Meta	92.5	91	89.5	92	90.5	91.1	256K
14	Claude Sonnet 4.6 Multimodal	Anthropic	91.5	94	85.2	90.8	93.5	91	200K
15	Qwen 3 Max Multimodal	Alibaba	91.5	91	88.5	90.5	90	90.3	256K
16	Gemini 3 Pro Multimodal	Google	91	90.5	87.8	91.5	90	90.2	1M
17	Grok 3 Multimodal	xAI	91.2	90.5	85	91.5	90	89.6	128K
18	OpenAI o3-mini	OpenAI	86.9	90	90.2	90.5	89.8	89.5	200K
19	Claude Sonnet 4 Multimodal	Anthropic	89.5	93.9	80.1	89.8	93.5	89.4	200K
20	Qwen 3 235B-A22B Open Source	Alibaba	89.5	91.5	85	90.2	91	89.4	128K
21	DeepSeek V3.5 Open Source	DeepSeek	90.5	91	86.5	89	90	89.4	128K
22	OpenAI o1-mini	OpenAI	85.2	90	90	88.5	89.5	88.6	128K
23	Mistral Large 3 Multimodal	Mistral AI	90	89.5	84	89.5	89	88.4	256K
24	Claude 3.5 Sonnet Multimodal	Anthropic	88.7	93.7	78.3	87.6	93	88.3	200K
25	Llama 4 Maverick Open SourceMultimodal	Meta	89.2	91	82.5	88	90	88.1	1M
26	Grok 3.5	xAI	89.5	89	84.5	89	88.5	88.1	128K
27	GPT-4.5 Multimodal	OpenAI	90.8	88.5	81.2	91.3	88	88	128K
28	GPT-4o Multimodal	OpenAI	88.7	90.2	76.6	86.4	89.5	86.3	128K
29	DeepSeek V3 Open Source	DeepSeek	87.1	89	82	84.5	88	86.1	128K
30	Llama 4 Scout Open SourceMultimodal	Meta	87.5	89.2	79	85.8	88	85.9	10M
31	Qwen 3 Plus Multimodal	Alibaba	87.5	87	82	86	86.5	85.8	128K
32	Ernie 5.0 Multimodal	Baidu	88	85	82.5	87.5	84	85.4	256K
33	Gemini 2.5 Flash Multimodal	Google	87	88.5	78.5	85.5	87	85.3	1M
34	GPT-5.2 Mini Multimodal	OpenAI	86.5	89.2	78.4	84	88.6	85.3	256K
35	Llama 3.1 405B Open Source	Meta	88.6	89	73.8	85	88.5	85	128K
36	Llama 3.3 70B Open Source	Meta	86	88.4	77	83.5	87	84.4	128K
37	Jamba 2	AI21 Labs	87	84.5	79	86	83.5	84	512K
38	Qwen 2.5 72B Open Source	Alibaba	85.3	86.4	80	82.5	85.5	83.9	128K
39	Gemini 3 Flash Multimodal	Google	85.5	86	78.5	84	85.5	83.9	1M
40	Yi-Lightning 2	01.AI	86	84.5	80	85	83.5	83.8	128K
41	Pixtral Large 2 Multimodal	Mistral AI	86	84	78.5	85.5	83.5	83.5	128K
42	Claude Haiku 4.5 Multimodal	Anthropic	85	88.5	74.2	82.5	87	83.4	200K
43	Grok 2 Multimodal	xAI	87.5	85	76	83.2	84.5	83.2	128K
44	Reka Core 2 Multimodal	Reka	86.5	83.5	78	85.5	82.5	83.2	128K
45	DeepSeek Coder V3 Open Source	DeepSeek	78.5	94	72	76.5	93.5	82.9	128K
46	Gemini 2.0 Flash Multimodal	Google	85.8	86.5	73.4	82.1	85	82.6	1M
47	Claude 3 Opus Multimodal	Anthropic	86.8	84.9	72	85	84.5	82.6	200K
48	Command R+ 2	Cohere	86.5	83	76	85.5	82	82.6	256K
49	WizardLM 3 Open Source	Microsoft	82.5	85	78	82	84.5	82.4	128K
50	Phi-4 Open Source	Microsoft	84.8	82.6	80.4	81	82	82.2	16K
51	Nemotron-4 340B Open Source	NVIDIA	85.5	82	78.5	84	81	82.2	128K
52	QwQ-32B-Preview Open Source	Alibaba	79.5	80	85.5	87	78.5	82.1	32K
53	Mistral Large 2	Mistral AI	84	84.5	74.5	82.8	84	82	128K
54	Falcon 3 180B Open Source	TII	85	82.5	77	84	81.5	82	128K
55	DBRX 2 Open Source	Databricks	84	83.5	75.5	83	82.5	81.7	128K
56	Claude 3.5 Haiku Multimodal	Anthropic	84	88.1	69.3	78.2	86.5	81.2	200K
57	HyperCLOVA X 2 Multimodal	Naver	85	80.5	76.5	84	79.5	81.1	128K
58	GPT-4o Mini Multimodal	OpenAI	82	87	70.2	78.5	85.3	80.6	128K
59	Gemini 1.5 Pro Multimodal	Google	85.9	84.1	67.7	82	83.5	80.6	2M
60	Qwen 3 Turbo Open Source	Alibaba	82	83.5	75	80.5	82	80.6	128K
61	Mistral Medium 3	Mistral AI	83.5	83	71.5	81	82.5	80.3	128K
62	Codestral 2	Mistral AI	72	93.5	68	74	93	80.1	256K
63	Qwen 2.5 Coder 32B Open Source	Alibaba	74.2	92.7	68.5	72	92	79.9	128K
64	Nous Hermes 3 Open Source	Nous Research	80.5	82	73	79.5	81	79.2	128K
65	Inflection Pi-3	Inflection AI	84	78.5	73	83	77	79.1	128K
66	Arctic 2 Open Source	Snowflake	81	80.5	74	80	79.5	79	128K
67	DeepSeek Coder V2 Open Source	DeepSeek	71	90.2	73.5	69	89.5	78.6	128K
68	Nemotron 70B Open Source	NVIDIA	83.5	80	68.5	80.5	79.5	78.4	128K
69	Command A	Cohere	82.8	80.5	68	80.2	79.5	78.2	256K
70	Llama 3.1 70B Open Source	Meta	82	80.5	68	79	80	77.9	128K
71	Dolphin 3 Open Source	Cognitive Computations	79	80.5	71	78	79.5	77.6	128K
72	GLM-4-Plus Multimodal	Zhipu AI	82	78	70.5	79	77.5	77.4	128K
73	Gemma 3 27B Open SourceMultimodal	Google	80.5	80	68	78.2	79	77.1	128K
74	Mistral Small 3 Open Source	Mistral AI	80.5	81	66.5	76.3	80	76.9	128K
75	Yi-Lightning	01.AI	82	78.5	68	78	77.5	76.8	16K
76	MiniMax-01 Open Source	MiniMax	82.5	78	66.5	79	77.5	76.7	4M
77	Codestral	Mistral AI	70.5	90	60	68.5	91	76	256K
78	Falcon 3 40B Open Source	TII	78.5	76	70	77	75.5	75.4	64K
79	Gemini 2.0 Flash Lite Multimodal	Google	80.2	78.5	65	74.3	77.8	75.2	1M
80	Solar 2 Pro Multimodal	Upstage	77.5	76	70.5	76.5	75	75.1	64K
81	Amazon Nova Pro Multimodal	Amazon	80.2	77.5	64	77	76.5	75	300K
82	Command R 2	Cohere	78	76.5	68.5	77	75	75	128K
83	Mixtral 8x22B Open Source	Mistral AI	77.8	79	64.2	75.5	78	74.9	64K
84	CodeLlama 2 70B Open Source	Meta	68	86	62	68.5	85.5	74	128K
85	Jamba 1.5 Large Open Source	AI21 Labs	80	75.5	62.4	76.8	75	73.9	256K
86	Gemma 3 9B Open SourceMultimodal	Google	74.5	76	68	73	75.5	73.4	128K
87	Yi-34B-Chat Open Source	01.AI	76.5	74	66	74.5	73	72.8	32K
88	StarCoder 3 Open Source	BigCode	65	88.5	58	65	87.5	72.8	64K
89	Cohere Aya 3 Open Source	Cohere	78.5	72	66.5	76	71	72.8	64K
90	Qwen 2.5 7B Open Source	Alibaba	74.2	75.6	65	71.5	74	72.1	128K
91	BLOOM-3 Open Source	BigScience	76	70.5	65	74.5	69.5	71.1	64K
92	Phi-4 Mini Open Source	Microsoft	72.5	72	68.3	70	71.5	70.9	128K
93	WizardLM 2 8x22B Open Source	Microsoft	75.5	74	58	73.5	73.5	70.9	64K
94	Reka Core Multimodal	Reka AI	79	72.5	56	75	72	70.9	128K
95	Claude 3 Haiku Multimodal	Anthropic	75.2	75.9	57.3	71	74.5	70.8	200K
96	Gemma 2 27B Open Source	Google	75.2	73.5	58.3	72	73	70.4	8K
97	Command R+	Cohere	75.7	71.5	57.8	74	72	70.2	128K
98	InternLM 2.5 20B Open Source	Shanghai AI Lab	73.8	72	62	71	71.5	70.1	1M
99	DBRX Open Source	Databricks	73.7	70.1	54.2	69.5	69.8	67.5	32K
100	Amazon Nova Lite Multimodal	Amazon	72.5	68	55	68	67.5	66.2	300K
101	Jamba 1.5 Mini Open Source	AI21 Labs	72	68	52.5	68.5	67	65.6	256K
102	Aya Expanse 32B Open Source	Cohere	73	65	52	70	64	64.8	128K
103	Mistral Nemo Open Source	Mistral AI	68	67.5	52	65.5	66.5	63.9	128K
104	Falcon 180B Open Source	TII	70.5	62	45	65	60.5	60.6	2K
105	Snowflake Arctic Open Source	Snowflake	67	64.5	45.5	62	63.5	60.5	4K
106	Llama 3.1 8B Open Source	Meta	68.4	62.1	47.2	62	61.5	60.2	128K
107	StarCoder 2 15B Open Source	BigCode	52	73.5	35	48.5	75	56.8	16K

Try All These Models in One Place

Vincony gives you access to 400+ AI models — compare responses side-by-side, run AI debates, and find the best model for your task.

Visit Vincony.com