GPQA Diamond

Name: GPQA Diamond — AI Model Benchmark Results
Creator: Neura Market

448 graduate-level science questions (biology, chemistry, physics) designed to be hard for non-experts.

Model Rankings (246)

#	Model	Score	Provider	Source
1	Claude Mythos Preview🇺🇸	94.6	Anthropic	verified
2	GPT-5.6 Sol🇺🇸	94.6	OpenAI	verified
3	Gemini 3.1 Pro🇺🇸	94.3	Google	verified
4	Claude Opus 4.7🇺🇸	94.2	Anthropic	verified
5	GPT-5.5🇺🇸	93.6	OpenAI	verified
6	Claude Opus 4.8🇺🇸	93.6	Anthropic	verified
7	M Kimi K3🇨🇳	93.5	Moonshot AI	verified
8	GPT-5.2 Pro🇺🇸	93.2	OpenAI	verified
9	X Grok 4.5🇺🇸	93	xAI	verified
10	GPT-5.6 Terra🇺🇸	92.9	OpenAI	verified
11	GPT-5.4🇺🇸	92.8	OpenAI	verified
12	GPT-5.2🇺🇸	92.4	OpenAI	verified
13	Q Qwen3.7 Max🇨🇳	92.4	Alibaba Cloud / Qwen Team	verified
14	GPT-5.6 Luna🇺🇸	92.3	OpenAI	verified
15	Gemini 3 Pro🇺🇸	91.9	Google	verified
16	Claude Opus 4.6🇺🇸	91.3	Anthropic	verified
17	Z GLM-5.2🇨🇳	91.2	Zhipu AI	verified
18	M Kimi K2.6🇨🇳	90.5	Moonshot AI	verified
19	T Hy3🇨🇳	90.4	Tencent	verified
20	Q Qwen3.6 Plus🇨🇳	90.4	Alibaba Cloud / Qwen Team	verified
21	Gemini 3 Flash🇺🇸	90.4	Google	verified
22	Q Qwen3.7-Plus🇨🇳	90.3	Alibaba Cloud / Qwen Team	verified
23	DeepSeek-V4-Pro-Max🇨🇳	90.1	DeepSeek	verified
24	Claude Sonnet 4.6🇺🇸	89.9	Anthropic	verified
25	M Muse Spark🇺🇸	89.5	Meta	verified
26	B Seed 2.0 Pro🇨🇳	88.9	ByteDance	verified
27	Q Qwen3.5-397B-A17B🇨🇳	88.4	Alibaba Cloud / Qwen Team	verified
28	X Grok-4 Heavy🇺🇸	88.4	xAI	verified
29	GPT-5.1 Thinking🇺🇸	88.1	OpenAI	verified
30	GPT-5 Medium🇺🇸	88.1	OpenAI	verified
31	DeepSeek-V4-Flash-Max🇨🇳	88.1	DeepSeek	verified
32	GPT-5.1 Instant🇺🇸	88.1	OpenAI	verified
33	GPT-5.1 High🇺🇸	88.1	OpenAI	verified
34	GPT-5.1🇺🇸	88.1	OpenAI	verified
35	GPT-5.4 mini🇺🇸	88	OpenAI	verified
36	Q Qwen3.6-27B🇨🇳	87.8	Alibaba Cloud / Qwen Team	verified
37	M Kimi K2.5🇨🇳	87.6	Moonshot AI	verified
38	X Grok-4🇺🇸	87.5	xAI	verified
39	GPT-5 High🇺🇸	87.3	OpenAI	verified
40	Nemotron 3 Ultra (550B A55B)🇺🇸	87	NVIDIA	verified
41	Claude Opus 4.5🇺🇸	87	Anthropic	verified
42	Gemini 3.1 Flash-Lite🇺🇸	86.9	Google	verified
43	Q Qwen3.5-122B-A10B🇨🇳	86.6	Alibaba Cloud / Qwen Team	verified
44	Gemini 2.5 Pro Preview 06-05🇺🇸	86.4	Google	verified
45	Z GLM-5.1🇨🇳	86.2	Zhipu AI	verified
46	Q Qwen3.6-35B-A3B🇨🇳	86	Alibaba Cloud / Qwen Team	verified
47	Z GLM-4.7🇨🇳	85.7	Zhipu AI	verified
48	GPT-5🇺🇸	85.7	OpenAI	verified
49	X Grok 4 Fast🇺🇸	85.7	xAI	verified
50	GPT-5.5 Instant🇺🇸	85.6	OpenAI	verified
51	Q Qwen3.5-27B🇨🇳	85.5	Alibaba Cloud / Qwen Team	verified
52	B Seed 2.0 Lite🇨🇳	85.1	ByteDance	verified
53	ERNIE 5.0🇨🇳	85	Baidu	verified
54	Claude 3.7 Sonnet🇺🇸	84.8	Anthropic	verified
55	X Grok-3🇺🇸	84.6	xAI	verified
56	MAI-Code-1-Flash🇺🇸	84.6	Microsoft	verified
57	M Kimi K2-Thinking-0905🇨🇳	84.5	Moonshot AI	verified
58	Gemma 4 31B🇺🇸	84.3	Google	verified
59	MAI-Thinking-1🇺🇸	84.2	Microsoft	verified
60	Q Qwen3.5-35B-A3B🇨🇳	84.2	Alibaba Cloud / Qwen Team	verified
61	ChatGPT-4o Latest🇺🇸	84	OpenAI	verified
62	X Grok-3 Mini🇺🇸	84	xAI	verified
63	X MiMo-V2-Flash🇨🇳	83.7	Xiaomi	verified
64	Claude Sonnet 4.5🇺🇸	83.4	Anthropic	verified
65	o3🇺🇸	83.3	OpenAI	verified
66	Gemini 2.5 Pro🇺🇸	83	Google	verified
67	GPT-5.4 nano🇺🇸	82.8	OpenAI	verified
68	Gemini 2.5 Flash🇺🇸	82.8	Google	verified
69	Nemotron 3 Super (120B A12B)🇺🇸	82.7	NVIDIA	verified
70	DeepSeek-V3.2🇨🇳	82.4	DeepSeek	verified
71	DeepSeek-V3.2 (Thinking)🇨🇳	82.4	DeepSeek	verified
72	GPT-5 mini🇺🇸	82.3	OpenAI	verified
73	Gemma 4 26B-A4B🇺🇸	82.3	Google	verified
74	Q Qwen3.5-9B🇨🇳	81.7	Alibaba Cloud / Qwen Team	verified
75	M LongCat-Flash-Thinking🇨🇳	81.5	Meituan	verified
76	Nova 2 Pro🇺🇸	81.4	Amazon	verified
77	o4-mini🇺🇸	81.4	OpenAI	verified
78	Q Qwen3-235B-A22B-Thinking-2507🇨🇳	81.1	Alibaba Cloud / Qwen Team	verified
79	DeepSeek-R1-0528🇨🇳	81	DeepSeek	verified
80	Z GLM-4.6🇨🇳	81	Zhipu AI	verified
81	MiniMax M2.1🇨🇳	81	MiniMax	verified
82	Claude Opus 4.1🇺🇸	80.9	Anthropic	verified
83	GPT OSS 120B High🇺🇸	80.9	OpenAI	verified
84	M LongCat-Flash-Thinking-2601🇨🇳	80.5	Meituan	verified
85	GPT OSS 120B🇺🇸	80.1	OpenAI	verified
86	DeepSeek-V3.2-Exp🇨🇳	79.9	DeepSeek	verified
87	Nova 2 Lite🇺🇸	79.6	Amazon	verified
88	Claude Opus 4🇺🇸	79.6	Anthropic	verified
89	Z GLM-4.5🇨🇳	79.1	Zhipu AI	verified
90	o1-pro🇺🇸	79	OpenAI	verified
91	Gemma 4 12B🇺🇸	78.8	Google	verified
92	S Sarvam-105B🇮🇳	78.7	Sarvam AI	verified
93	o1🇺🇸	78	OpenAI
94	MiniMax M2🇨🇳	78	MiniMax	verified
95	o1🇺🇸	78	OpenAI	verified
96	Q Qwen3-235B-A22B-Instruct-2507🇨🇳	77.5	Alibaba Cloud / Qwen Team	verified
97	o3-mini🇺🇸	77.2	OpenAI	verified
98	Q Qwen3-Next-80B-A3B-Thinking🇨🇳	77.2	Alibaba Cloud / Qwen Team	verified
99	Q Qwen3.5-4B🇨🇳	76.2	Alibaba Cloud / Qwen Team	verified
100	Llama 3.1 Nemotron Ultra 253B v1🇺🇸	76.0	NVIDIA	verified
101	M Kimi K2 0905🇨🇳	75.8	Moonshot AI	verified
102	Claude Sonnet 4🇺🇸	75.4	Anthropic	verified
103	Z GLM-4.7-Flash🇨🇳	75.2	Zhipu AI	verified
104	M Kimi K2 Instruct🇨🇳	75.1	Moonshot AI	verified
105	M Kimi K2-Instruct-0905🇨🇳	75.1	Moonshot AI	verified
106	Nemotron 3 Nano (30B A3B)🇺🇸	75	NVIDIA	verified
107	Z GLM-4.5-Air🇨🇳	75	Zhipu AI	verified
108	DeepSeek-V3.1🇨🇳	74.9	DeepSeek	verified
109	Claude Opus 4🇺🇸	74.8	Anthropic
110	Q Qwen3 VL 30B A3B Thinking🇨🇳	74.4	Alibaba Cloud / Qwen Team	verified
111	Gemini 2.0 Flash Thinking🇺🇸	74.2	Google	verified
112	GPT OSS 20B High🇺🇸	74.2	OpenAI	verified
113	I Mercury 2🇺🇸	74	Inception	verified
114	ERNIE 4.5🇨🇳	74	Baidu	verified
115	DeepSeek R1 Zero🇨🇳	73.3	DeepSeek	verified
116	o1-preview🇺🇸	73.3	OpenAI	verified
117	M LongCat-Flash-Chat🇨🇳	73.2	Meituan	verified
118	DiffusionGemma 26B-A4B🇺🇸	73.2	Google	verified
119	Q Qwen3 VL 32B Thinking🇨🇳	73.1	Alibaba Cloud / Qwen Team	verified
120	Claude Haiku 4.5🇺🇸	73	Anthropic	verified
121	Q Qwen3-Next-80B-A3B-Instruct🇨🇳	72.9	Alibaba Cloud / Qwen Team	verified
122	GPT OSS 20B🇺🇸	71.5	OpenAI	verified
123	Ministral 3 (14B Reasoning 2512)🇫🇷	71.2	Mistral AI	verified
124	Mistral Small 4🇫🇷	71.2	Mistral AI	verified
125	GPT-5 nano🇺🇸	71.2	OpenAI	verified
126	Magistral Medium🇫🇷	70.8	Mistral AI	verified
127	Claude Sonnet 4🇺🇸	70.5	Anthropic
128	Q Qwen3 VL 30B A3B Instruct🇨🇳	70.4	Alibaba Cloud / Qwen Team	verified
129	GPT-4o🇺🇸	70.1	OpenAI	verified
130	MiniMax M1 80K🇨🇳	70	MiniMax	verified
131	Q Qwen3 VL 8B Thinking🇨🇳	69.9	Alibaba Cloud / Qwen Team	verified
132	M Llama 4 Maverick🇺🇸	69.8	Meta	verified
133	GPT-4.5🇺🇸	69.5	OpenAI	verified
134	MiniMax M1 40K🇨🇳	69.2	MiniMax	verified
135	Q Qwen3 VL 32B Instruct🇨🇳	68.9	Alibaba Cloud / Qwen Team	verified
136	Phi 4 Reasoning Plus🇺🇸	68.9	Microsoft	verified
137	DeepSeek-V3 0324🇨🇳	68.4	DeepSeek	verified
138	Magistral Small 2506🇫🇷	68.2	Mistral AI	verified
139	Claude 3.5 Sonnet🇺🇸	67.2	Anthropic	verified
140	Ministral 3 (8B Reasoning 2512)🇫🇷	66.8	Mistral AI	verified
141	M LongCat-Flash-Lite🇨🇳	66.8	Meituan	verified
142	X MiMo-V2.5-Pro🇨🇳	66.7	Xiaomi	verified
143	Llama-3.3 Nemotron Super 49B v1🇺🇸	66.7	NVIDIA	verified
144	S Sarvam-30B🇮🇳	66.5	Sarvam AI	verified
145	GPT-4.1🇺🇸	66.3	OpenAI	verified
146	N Hermes 3 70B🇺🇸	66.1	Nous Research	verified
147	Phi 4 Reasoning🇺🇸	65.8	Microsoft	verified
148	Q Qwen3 30B A3B🇨🇳	65.8	Alibaba Cloud / Qwen Team	verified
149	DeepSeek R1 Distill Llama 70B🇨🇳	65.2	DeepSeek	verified
150	Q QwQ-32B🇨🇳	65.2	Alibaba Cloud / Qwen Team	verified
151	Q QwQ-32B-Preview🇨🇳	65.2	Alibaba Cloud / Qwen Team	verified
152	Claude 3.5 Sonnet🇺🇸	65	Anthropic
153	GPT-4.1 mini🇺🇸	65	OpenAI	verified
154	Gemini 2.5 Flash-Lite🇺🇸	64.6	Google	verified
155	Q Qwen3 VL 4B Thinking🇨🇳	64.1	Alibaba Cloud / Qwen Team	verified
156	Nemotron Nano 9B v2🇺🇸	64	NVIDIA	verified
157	DeepSeek R1 Distill Qwen 32B🇨🇳	62.1	DeepSeek	verified
158	Gemini 2.0 Flash🇺🇸	62.1	Google	verified
159	Q Qwen3 Max🇨🇳	62	Alibaba Cloud / Qwen Team	verified
160	o1-mini🇺🇸	60	OpenAI	verified
161	Claude 3.5 Sonnet🇺🇸	59.4	Anthropic	verified
162	DeepSeek-V3🇨🇳	59.1	DeepSeek	verified
163	DeepSeek R1 Distill Qwen 14B🇨🇳	59.1	DeepSeek	verified
164	Gemini 1.5 Pro🇺🇸	59.1	Google	verified
165	Gemma 4 E4B🇺🇸	58.6	Google	verified
166	M Llama 4 Scout🇺🇸	57.2	Meta	verified
167	Phi 4🇺🇸	56.1	Microsoft	verified
168	X Grok-2🇺🇸	56	xAI	verified
169	Llama 3.1 Nemotron Nano 8B V1🇺🇸	54.1	NVIDIA	verified
170	GPT-4o🇺🇸	53.6	OpenAI
171	GPT-4o🇺🇸	53.6	OpenAI	verified
172	Min istral 3 (3B Reasoning 2512)🇫🇷	53.4	Mistral AI	verified
173	Phi 4 Mini Reasoning🇺🇸	52	Microsoft	verified
174	Q Qwen3.5-2B🇨🇳	51.6	Alibaba Cloud / Qwen Team	verified
175	Gemini 2.0 Flash-Lite🇺🇸	51.5	Google	verified
176	Llama 3.1 405B🇺🇸	51.1	Meta
177	Gemini 1.5 Flash🇺🇸	51	Google	verified
178	X Grok-2 mini🇺🇸	51	xAI	verified
179	M Llama 3.1 405B Instruct🇺🇸	50.7	Meta	verified
180	M Llama 3.3 70B Instruct🇺🇸	50.5	Meta	verified
181	Claude 3 Opus🇺🇸	50.4	Anthropic
182	Claude 3 Opus🇺🇸	50.4	Anthropic	verified
183	GPT-4.1 nano🇺🇸	50.3	OpenAI	verified
184	Llama 3.3 70B🇺🇸	49.5	Meta
185	Q Qwen2.5 32B Instruct🇨🇳	49.5	Alibaba Cloud / Qwen Team	verified
186	GPT-4 Turbo🇺🇸	49.3	OpenAI
187	DeepSeek R1 Distill Qwen 7B🇨🇳	49.1	DeepSeek	verified
188	Q Qwen2.5 72B Instruct🇨🇳	49	Alibaba Cloud / Qwen Team	verified
189	DeepSeek R1 Distill Llama 8B🇨🇳	49	DeepSeek	verified
190	Qwen 2.5 72B🇨🇳	49	Alibaba
191	Mistral Large 2🇫🇷	48.2	Mistral
192	M Kimi K2 Base🇨🇳	48.1	Moonshot AI	verified
193	GPT-4 Turbo🇺🇸	48	OpenAI	verified
194	Q Qwen3 235B A22B🇨🇳	47.5	Alibaba Cloud / Qwen Team	verified
195	Nova Pro🇺🇸	46.9	Amazon	verified
196	M Llama 3.2 90B Instruct🇺🇸	46.7	Meta	verified
197	Llama 3.1 70B🇺🇸	46.7	Meta
198	Mistral Small 3.2 24B Instruct🇫🇷	46.1	Mistral AI	verified
199	Q Qwen2.5 VL 32B Instruct🇨🇳	46	Alibaba Cloud / Qwen Team	verified
200	Mistral Small 3.1 24B Instruct🇫🇷	46.0	Mistral AI	verified
201	Q Qwen2.5 14B Instruct🇨🇳	45.5	Alibaba Cloud / Qwen Team	verified
202	Mistral Small 3 24B Instruct🇫🇷	45.3	Mistral AI	verified
203	Mistral Large 3 (675B Instruct 2512)🇫🇷	43.9	Mistral AI	verified
204	Mistral Large 3 (675B Instruct 2512 NVFP4)🇫🇷	43.9	Mistral AI	verified
205	Mistral Large 3 (675B Base)🇫🇷	43.9	Mistral AI	verified
206	Mistral Large 3 (675B Instruct 2512 Eagle)🇫🇷	43.9	Mistral AI	verified
207	Gemma 4 E2B🇺🇸	43.4	Google	verified
208	Q Qwen2 72B Instruct🇨🇳	42.4	Alibaba Cloud / Qwen Team	verified
209	Gemma 3 27B🇺🇸	42.4	Google	verified
210	Nova Lite🇺🇸	42	Amazon	verified
211	M Llama 3.1 70B Instruct🇺🇸	41.7	Meta	verified
212	Claude 3.5 Haiku🇺🇸	41.6	Anthropic
213	Claude 3.5 Haiku🇺🇸	41.6	Anthropic	verified
214	Gemma 3 12B🇺🇸	40.9	Google	verified
215	Gemini Diffusion🇺🇸	40.4	Google	verified
216	Claude 3 Sonnet🇺🇸	40.4	Anthropic	verified
217	GPT-4o mini🇺🇸	40.2	OpenAI
218	GPT-4o mini🇺🇸	40.2	OpenAI	verified
219	Nova Micro🇺🇸	40	Amazon	verified
220	Gemini 1.5 Flash 8B🇺🇸	38.4	Google	verified
221	Mistral Small 3.1 24B Base🇫🇷	37.5	Mistral AI	verified
222	Jamba 1.5 Large🇮🇱	36.9	AI21 Labs	verified
223	Phi-3.5-MoE-instruct🇺🇸	36.8	Microsoft	verified
224	Q Qwen2.5 7B Instruct🇨🇳	36.4	Alibaba Cloud / Qwen Team	verified
225	X Grok-1.5🇺🇸	35.9	xAI	verified
226	GPT-4🇺🇸	35.7	OpenAI	verified
227	Mistral Small 3 24B Base🇫🇷	34.4	Mistral AI	verified
228	DeepSeek R1 Distill Qwen 1.5B🇨🇳	33.8	DeepSeek	verified
229	Claude 3 Haiku🇺🇸	33.3	Anthropic	verified
230	M Llama 3.2 11B Instruct🇺🇸	32.8	Meta	verified
231	M Llama 3.2 3B Instruct🇺🇸	32.8	Meta	verified
232	Jamba 1.5 Mini🇮🇱	32.3	AI21 Labs	verified
233	Gemma 3 4B🇺🇸	30.8	Google	verified
234	GPT-3.5 Turbo🇺🇸	30.8	OpenAI	verified
235	Q Qwen2.5-Omni-7B🇨🇳	30.8	Alibaba Cloud / Qwen Team	verified
236	Phi-3.5-mini-instruct🇺🇸	30.4	Microsoft	verified
237	M Llama 3.1 8B Instruct🇺🇸	30.4	Meta	verified
238	Gemini 1.0 Pro🇺🇸	27.9	Google	verified
239	Q Qwen2 7B Instruct🇨🇳	25.3	Alibaba Cloud / Qwen Team	verified
240	Phi 4 Mini🇺🇸	25.2	Microsoft	verified
241	Gemma 3n E2B Instructed🇺🇸	24.8	Google	verified
242	Gemma 3n E2B Instructed LiteRT (Preview)🇺🇸	24.8	Google	verified
243	Gemma 3n E4B Instructed🇺🇸	23.7	Google	verified
244	Gemma 3n E4B Instructed LiteRT Preview🇺🇸	23.7	Google	verified
245	Gemma 3 1B🇺🇸	19.2	Google	verified
246	Q Qwen3.5-0.8B🇨🇳	11.9	Alibaba Cloud / Qwen Team	verified

AI Model Benchmarks

GPQA Diamond

Model Rankings (246)