aa Gpqa Diamond

122 models evaluated

#	Model	Provider	Type	Score
1	Gemini 3.1 Pro	Google	Closed	94.1%
2	GPT-5.5	OpenAI	Closed	93.5%
3	Qwen3.7 Max	Alibaba	Closed	92.3%
4	Gemini 3.5 Flash	Google	Closed	92.2%
5	GPT-5.4	OpenAI	Closed	92.0%
6	GPT-5.3 Codex	OpenAI	Closed	91.5%
7	Claude Opus 4.7 (Adaptive)	Anthropic	Closed	91.4%
8	Kimi K2.6	Moonshot AI	Open	91.1%
9	Gemini 3 Pro	Google	Closed	90.8%
10	DeepSeek V4 Pro (High)	DeepSeek	Open	90.5%
11	GPT-5.2	OpenAI	Closed	90.3%
12	Grok 4.3	xAI	Closed	90.1%
13	GPT-5.2-Codex	OpenAI	Closed	89.9%
14	Claude Opus 4.6 (Adaptive)	Anthropic	Closed	89.6%
15	DeepSeek V4 Flash (Max)	DeepSeek	Open	89.4%
16	Qwen3.5 397B (Reasoning)	Alibaba	Open	89.3%
17	DeepSeek V4 Pro (Max)	DeepSeek	Open	88.8%
18	Qwen 3.6 Max (preview)	Alibaba	Closed	88.8%
19	Claude Opus 4.7	Anthropic	Closed	88.5%
20	Muse Spark	Meta	Closed	88.4%
21	Qwen3.6 Plus	Alibaba	Closed	88.2%
22	Kimi K2.5 (Reasoning)	Moonshot AI	Closed	87.9%
23	Kimi K2.5	Moonshot AI	Open	87.9%
24	Grok 4	xAI	Closed	87.7%
25	GPT-5.4 mini	OpenAI	Closed	87.5%
26	MiniMax M2.7	MiniMax	Open	87.4%
27	GPT-5.1	OpenAI	Closed	87.3%
28	MiMo-V2-Pro	Xiaomi	Closed	87.0%
29	GLM-5.1	Z.AI	Open	86.8%
30	DeepSeek V4 Flash (High)	DeepSeek	Open	86.7%
31	Hy3 Preview	Tencent	Open	86.7%
32	MiMo-V2.5-Pro	Xiaomi	Closed	86.6%
33	Claude Opus 4.5 Thinking	Anthropic	Closed	86.6%
34	Qwen3.5 397B	Alibaba	Open	86.1%
35	GPT-5.1-Codex-Max	OpenAI	Closed	86.0%
36	GPT-5.1-Codex	OpenAI	Closed	86.0%
37	GLM-4.7	Z.AI	Open	85.9%
38	Qwen3.5-27B	Alibaba	Open	85.8%
39	Qwen3.5-122B-A10B	Alibaba	Open	85.7%
40	Gemma 4 31B	Google	Open	85.7%
41	GPT-5 (high)	OpenAI	Closed	85.4%
42	Grok 4.1 Fast (Reasoning)	xAI	Closed	85.3%
43	GLM-5-Turbo	Z.AI	Closed	84.7%
44	Grok 4 Fast (Reasoning)	xAI	Closed	84.7%
45	o3-pro	OpenAI	Closed	84.5%
46	Qwen3.5-35B-A3B	Alibaba	Open	84.5%
47	Gemini 2.5 Pro	Google	Closed	84.4%
48	Qwen3.6-27B	Alibaba	Open	84.2%
49	GPT-5 (medium)	OpenAI	Closed	84.2%
50	Qwen3.6-35B-A3B	Alibaba	Open	84.1%
51	Claude Opus 4.6	Anthropic	Closed	84.0%
52	MiMo-V2-Omni	Xiaomi	Closed	82.8%
53	o3	OpenAI	Closed	82.7%
54	Gemini 3.1 Flash-Lite	Google	Closed	82.2%
55	GLM-5	Z.AI	Open	82.0%
56	GPT-5.4 nano	OpenAI	Closed	81.7%
57	DeepSeek-R1	DeepSeek	Open	81.3%
58	Gemini 3 Flash	Google	Closed	81.2%
59	Claude Opus 4.5	Anthropic	Closed	81.0%
60	Claude 4.1 Opus Thinking	Anthropic	Closed	80.9%
61	GLM-5V-Turbo	Z.AI	Closed	80.9%
62	Claude Sonnet 4.6	Anthropic	Closed	79.9%
63	Gemma 4 26B A4B	Google	Open	79.2%
64	K-Exaone	LG AI Research	Closed	78.3%
65	GPT-OSS 120B	OpenAI	Open	78.2%
66	DeepSeek V3.1 (Reasoning)	DeepSeek	Open	77.9%
67	Mistral Small 4 (Reasoning)	Mistral	Open	76.9%
68	Mistral Small 4	Mistral	Open	76.9%
69	Kimi K2	Moonshot AI	Closed	76.6%
70	Qwen3 Max	Alibaba	Closed	76.4%
71	Command A+	Cohere	Open	76.1%
72	Trinity-Large-Preview	Arcee AI	Open	75.2%
73	Trinity-Large-Thinking	Arcee AI	Open	75.2%
74	DeepSeek V3.2	DeepSeek	Open	75.1%
75	o3-mini	OpenAI	Closed	74.8%
76	Mistral Medium 3.5 128B	Mistral	Open	74.8%
77	o1	OpenAI	Closed	74.7%
78	Sarvam 105B	Sarvam	Open	73.8%
79	DeepSeek V3.1	DeepSeek	Open	73.5%
80	GLM-4.5-Air	Z.AI	Closed	73.3%
81	Nemotron Ultra 253B	NVIDIA	Open	72.8%
82	Grok Code Fast 1	xAI	Closed	72.7%
83	GPT-OSS 20B	OpenAI	Open	68.8%
84	Claude 4 Sonnet	Anthropic	Closed	68.3%
85	Gemini 2.5 Flash	Google	Closed	68.3%
86	Mistral Large 3	Mistral	Closed	68.0%
87	Llama 4 Maverick	Meta	Open	67.1%
88	GPT-4.1	OpenAI	Closed	66.6%
89	GPT-4.1 mini	OpenAI	Closed	66.4%
90	MiMo-V2-Flash	Xiaomi	Open	65.6%
91	Grok 4.1 Fast	xAI	Closed	63.7%
92	Sarvam 30B	Sarvam	Open	63.3%
93	GLM-4.6	Z.AI	Open	63.2%
94	Exaone 4.0 32B	LG AI Research	Open	62.8%
95	DeepSeek R1 Distill Qwen 32B	DeepSeek	Open	61.5%
96	Ling 2.6 Flash	InclusionAI	Open	59.3%
97	Gemini 1.5 Pro	Google	Closed	58.9%
98	Llama 4 Scout	Meta	Open	58.7%
99	Mistral Medium 3	Mistral	Closed	57.8%
100	Gemma 4 E4B	Google	Open	57.6%
101	Phi-4	Microsoft	Open	57.5%
102	Solar Pro 2	Upstage	Closed	56.1%
103	DeepSeek V3	DeepSeek	Open	55.7%
104	GPT-4o	OpenAI	Closed	54.3%
105	Llama 3.1 405B	Meta	Open	51.5%
106	GPT-4.1 nano	OpenAI	Closed	51.2%
107	Nova Pro	Amazon	Closed	49.9%
108	Claude 3 Opus	Anthropic	Closed	48.9%
109	Mistral Large 2	Mistral	Closed	48.6%
110	Nemotron 3 Nano Omni 30B A3B	NVIDIA	Open	46.9%
111	Gemma 4 E2B	Google	Open	43.3%
112	Gemma 3 27B	Google	Open	42.8%
113	GPT-4o mini	OpenAI	Closed	42.6%
114	Exaone 4.0 1.2B	LG AI Research	Open	42.4%
115	Qwen2.5 Coder 32B Instruct	Alibaba	Open	41.7%
116	Nemotron 3 Nano 30B	NVIDIA	Open	39.9%
117	Claude 3 Haiku	Anthropic	Closed	37.4%
118	Granite-4.0-1B	IBM	Open	28.1%
119	Gemini 1.0 Pro	Google	Closed	27.7%
120	Granite-4.0-H-1B	IBM	Open	26.3%
121	Granite-4.0-350M	IBM	Open	26.1%
122	Granite-4.0-H-350M	IBM	Open	25.7%