How is this ranking computed?

It is the same family composite as the leaderboard. Each benchmark is weighted by how much of its spread survives its own measurement error.

Best AI for reasoning

Ranked by the reasoning composite: GPQA Diamond, EBR-bench, Chess Puzzles and Mystery Game Puzzles. Full leaderboard.

Rank	Model	Lab	Composite	Evidence	Out $/M
1	GPT-5.6 Sol	OpenAI	100th	2 bench · 1 src	$10.00
2	GPT-6 Astra	OpenAI	100th	11 bench · 2 src	$50.00
3	GPT-5.5 Pro pre-release	OpenAI	99th	5 bench · 1 src	—
4	Claude Fable 5.1	Anthropic	99th	10 bench · 2 src	$50.00
5	Gemini 3.8 Flash	Google	98th	8 bench · 2 src	$3.75
6	GPT-5.4 Pro	OpenAI	98th	6 bench · 1 src	$180.00
7	DeepSeek V4 Pro 0813	DeepSeek	97th	8 bench · 2 src	$4.0077
8	Gemini 3.7 Flash	Google	97th	8 bench · 2 src	$3.75
9	Qwen3.8 Max (0902)	Alibaba	97th	6 bench · 1 src	—
10	GPT-5.5	OpenAI	96th	16 bench · 2 src	$30.00
11	Claude Fable 5	Anthropic	96th	14 bench · 2 src	$50.00
12	Claude Opus 5	Anthropic	95th	9 bench · 2 src	$25.00
13	GPT-5.6 Sol	OpenAI	95th	12 bench · 2 src	$10.00
14	Grok 4.5	xAI	94th	7 bench · 2 src	$6.00
15	Qwen3.8 Max	Alibaba	94th	8 bench · 2 src	$6.00
16	Muse Spark	Meta	94th	4 bench · 1 src	—
17	Gemini 3 Pro Preview	Google	93rd	6 bench · 2 src	$9.60
18	DeepSeek V4 Flash 0731	DeepSeek	93rd	8 bench · 2 src	$0.50
19	Gemini 3.1 Pro Preview	Google	92nd	12 bench · 2 src	$12.00
20	Grok 4.6	xAI	92nd	9 bench · 2 src	$6.00
21	Gemini 3.6 Flash	Google	92nd	8 bench · 2 src	$3.75
22	Gemini 3 Flash Preview	Google	91st	10 bench · 2 src	$3.00
23	Grok 4	xAI	91st	6 bench · 2 src	$15.00
24	Claude Sonnet 5	Anthropic	90th	8 bench · 2 src	$10.00
25	Gemini 2.5 Pro Exp	Google	90th	2 bench · 1 src	—
26	Grok 4.3 Beta	xAI	89th	6 bench · 1 src	—
27	GPT-5.6 Terra	OpenAI	89th	8 bench · 2 src	$12.00
28	Grok 4.20	xAI	89th	6 bench · 1 src	$6.00
29	Claude Opus 4.8	Anthropic	88th	13 bench · 2 src	$25.00
30	GLM-5.3	Zhipu	88th	8 bench · 2 src	$4.40
31	GPT-5.4	OpenAI	87th	13 bench · 2 src	$15.00
32	Qwen3.6 35B-A3B	Alibaba	87th	4 bench · 1 src	$2.052
33	o3	OpenAI	86th	11 bench · 3 src	$8.00
34	Inkling	Thinking Machines	86th	7 bench · 2 src	$4.05
35	Kimi K2.7 Code	Kimi	86th	7 bench · 2 src	$4.389
36	Qwen3.6 27B	Alibaba	85th	5 bench · 2 src	free
37	GLM-5.1	Zhipu	85th	8 bench · 1 src	$4.40
38	Qwen3.6 Max Preview	Alibaba	84th	7 bench · 1 src	free
39	GPT 5.2	OpenAI	84th	12 bench · 3 src	$14.00
40	Gemini 3.5 Flash	Google	83rd	14 bench · 3 src	$9.00
41	DeepSeek v4	DeepSeek	83rd	7 bench · 1 src	—
42	Gemini 2.5 Pro	Google	83rd	8 bench · 2 src	$10.00
43	Qwen3.6 Plus	Alibaba	82nd	9 bench · 2 src	$3.00
44	Kimi K2 Instruct	Kimi	82nd	7 bench · 3 src	free
45	Kimi K3	Kimi	81st	8 bench · 2 src	free
46	GPT-5.6 Luna	OpenAI	81st	8 bench · 2 src	$0.37
47	Claude Sonnet 4	Anthropic	80th	7 bench · 3 src	$15.00
48	Claude 3.7 Sonnet	Anthropic	80th	7 bench · 3 src	$15.00
49	DeepSeek-R1	DeepSeek	80th	4 bench · 1 src	—
50	Gemini 2.5 Pro Preview	Google	79th	4 bench · 1 src	—
51	Grok-3 mini	xAI	79th	5 bench · 1 src	$0.50
52	Gemini 3.1 Flash-Lite	Google	78th	4 bench · 1 src	$1.50
53	Qwen3.7 Plus	Alibaba	78th	4 bench · 1 src	$3.072
54	GPT 5.1	OpenAI	77th	8 bench · 2 src	$10.00
55	Kimi K2.5	Kimi	77th	7 bench · 2 src	$1.90
56	Claude Opus 4.7	Anthropic	77th	13 bench · 2 src	$25.00
57	Claude Opus 4.8	Anthropic	76th	3 bench · 1 src	—
58	GPT-5.6 Sol	OpenAI	76th	4 bench · 1 src	—
59	Qwen3.5 397B-A17B	Alibaba	75th	4 bench · 1 src	$2.65
60	Qwen3.5 Flash	Alibaba	75th	8 bench · 1 src	$0.40
61	Claude Opus 4	Anthropic	74th	7 bench · 2 src	$75.00
62	GLM 5	Zhipu	74th	6 bench · 2 src	$3.2647
63	Qwen3.5 Plus	Alibaba	74th	7 bench · 1 src	$2.40
64	Qwen3.7-Max	Alibaba	73rd	10 bench · 2 src	$9.00
65	GPT-5.5 Instant	OpenAI	73rd	5 bench · 1 src	$30.00
66	Qwen3.7 Flash	Alibaba	72nd	5 bench · 1 src	$0.13
67	Qwen3	Alibaba	72nd	3 bench · 1 src	—
68	Nemotron 3 Ultra	NVIDIA	71st	4 bench · 1 src	$3.00
69	Qwen3.5 397B-A17B	Alibaba	71st	5 bench · 1 src	—
70	Qwen3.5 9B	Alibaba	71st	3 bench · 1 src	—
71	DeepSeek R1	DeepSeek	70th	4 bench · 2 src	$1.70
72	GPT-4.5	OpenAI	70th	4 bench · 2 src	$150.00
73	Qwen3.6 Flash	Alibaba	69th	8 bench · 1 src	$1.16
74	Gemini 3.5 Flash-Lite	Google	69th	7 bench · 2 src	$2.50
75	GPT 5	OpenAI	68th	13 bench · 3 src	$10.00
76	DeepSeek-V3	DeepSeek	68th	4 bench · 1 src	—
77	Grok 3 Beta	xAI	68th	6 bench · 2 src	—
78	GLM-4.7	Zhipu	67th	6 bench · 1 src	$0.80
79	GPT-5.4 mini	OpenAI	67th	10 bench · 2 src	$4.50
80	Llama 4 Maverick Instruct	Meta	66th	6 bench · 3 src	$0.60
81	Claude Sonnet 4.5	Anthropic	66th	10 bench · 1 src	$15.00
82	Qwen3.6 35B-A3B	Alibaba	65th	5 bench · 1 src	—
83	Qwen3.5-35B-A3B	Alibaba	65th	3 bench · 1 src	$1.80
84	seed-oss-36b-instruct	—	65th	3 bench · 1 src	$0.57
85	Inkling Small	Thinking Machines	64th	7 bench · 1 src	$1.20
86	Claude Opus 4.5	Anthropic	64th	11 bench · 2 src	$25.00
87	GLM-5.2	Zhipu	63rd	10 bench · 2 src	$4.40
88	Kimi K2.6	Kimi	63rd	12 bench · 2 src	$4.00
89	GLM-5.3-Flash	Zhipu	62nd	7 bench · 2 src	$0.025
90	Gemini 2.0 Pro	Google	62nd	3 bench · 1 src	$4.998
91	MiniMax M3	MiniMax	62nd	5 bench · 2 src	$2.40
92	QwQ-Plus	Alibaba	61st	1 bench · 1 src	$0.574
93	Qwen3.7 Plus	Alibaba	61st	5 bench · 1 src	—
94	DeepSeek-V3.2-Exp	DeepSeek	60th	7 bench · 3 src	$0.326
95	GPT-5.5	OpenAI	60th	4 bench · 1 src	—
96	Claude Opus 4.5	Anthropic	59th	6 bench · 1 src	—
97	Claude Opus 4.6	Anthropic	59th	11 bench · 2 src	$25.00
98	Claude Sonnet 4.6	Anthropic	59th	8 bench · 2 src	$15.94
99	o1-2024-12-17	OpenAI	58th	9 bench · 3 src	$60.00
100	GPT-5.4 nano	OpenAI	58th	10 bench · 2 src	$1.25
101	Qwen3.7 Flash	Alibaba	57th	5 bench · 1 src	—
102	GPT-5.1	OpenAI	57th	5 bench · 1 src	—
103	Gemma 4 31B IT	Google	56th	4 bench · 1 src	$0.42
104	o4-mini	OpenAI	56th	12 bench · 3 src	$4.40
105	Qwen3.5 9B	Alibaba	56th	3 bench · 1 src	$0.15
106	Gemma 4 26B-A4B IT	Google	55th	3 bench · 1 src	$0.29
107	Qwen3-235B-A22B	Alibaba	55th	7 bench · 1 src	—
108	GPT 5 mini	OpenAI	54th	11 bench · 2 src	$2.00
109	Qwen3.6 Max Preview	Alibaba	54th	1 bench · 1 src	—
110	DeepSeek v4	DeepSeek	53rd	4 bench · 1 src	—
111	Qwen3.6 27B	Alibaba	53rd	5 bench · 1 src	—
112	o1-mini	OpenAI	53rd	5 bench · 2 src	$4.40
113	GLM-5.2	Zhipu	52nd	5 bench · 1 src	—
114	Gemini 2.0 Flash Thinking	Google	52nd	6 bench · 3 src	$0.42
115	GPT-5.4	OpenAI	51st	4 bench · 1 src	—
116	GPT-5.6 Terra	OpenAI	51st	4 bench · 1 src	—
117	Qwen3.6 Flash	Alibaba	50th	2 bench · 1 src	—
118	gpt-oss-120b	OpenAI	50th	6 bench · 3 src	$0.798
119	Claude Haiku 4.5	Anthropic	50th	7 bench · 1 src	$5.00
120	Claude Opus 4.1	Anthropic	49th	10 bench · 1 src	$75.00
121	GPT 5 nano	OpenAI	49th	11 bench · 2 src	$0.40
122	Mistral Medium 3	Mistral	48th	4 bench · 1 src	$2.00
123	GPT 4.1	OpenAI	48th	10 bench · 3 src	$8.00
124	GPT-5.2	OpenAI	47th	4 bench · 1 src	—
125	Qwen3.5 Plus	Alibaba	47th	1 bench · 1 src	—
126	Qwen3.5-122B-A10B	Alibaba	47th	1 bench · 1 src	$3.496
127	MiniMax M3	MiniMax	46th	4 bench · 1 src	—
128	O3 Mini	OpenAI	46th	12 bench · 3 src	$4.40
129	QwQ 32B	Alibaba	45th	4 bench · 2 src	$0.861
130	GPT-5.6 Luna	OpenAI	45th	5 bench · 1 src	—
131	Qwen3.5 Flash	Alibaba	44th	2 bench · 1 src	—
132	Claude Sonnet 5	Anthropic	44th	1 bench · 1 src	—
133	DeepSeek-V3	DeepSeek	44th	4 bench · 1 src	$0.77
134	Qwen3-30B-A3B	Alibaba	43rd	3 bench · 1 src	—
135	Qwen2.5-Max	Alibaba	43rd	4 bench · 1 src	$6.392
136	Magistral Small 1.0	Mistral	42nd	3 bench · 1 src	—
137	DeepSeek-R1-Distill-Llama-70B	DeepSeek	42nd	3 bench · 1 src	$0.99
138	Claude 3.5 Sonnet	Anthropic	42nd	5 bench · 1 src	—
139	Claude Opus 4.6	Anthropic	41st	4 bench · 1 src	—
140	GPT 4o	OpenAI	41st	9 bench · 3 src	$10.00
141	Claude 3.5 Sonnet	Anthropic	40th	5 bench · 1 src	—
142	DeepSeek-R1-Distill-Qwen-32B	DeepSeek	40th	3 bench · 1 src	$0.861
143	Grok-2	xAI	39th	4 bench · 1 src	$10.00
144	GPT 4.1 mini	OpenAI	39th	10 bench · 3 src	$1.60
145	Claude Sonnet 4.6	Anthropic	39th	2 bench · 1 src	—
146	Qwen3 Max	Alibaba	38th	7 bench · 1 src	$6.001
147	Qwen3 32B	Alibaba	38th	4 bench · 2 src	$0.55
148	Claude Sonnet 4.5	Anthropic	37th	8 bench · 1 src	—
149	GPT-5.4 mini	OpenAI	37th	5 bench · 1 src	—
150	Qwen3 8B	Alibaba	36th	3 bench · 1 src	$0.47
151	Qwen3-14B	Alibaba	36th	3 bench · 1 src	$0.24
152	Llama 4 Scout Instruct	Meta	36th	5 bench · 2 src	$0.46
153	Claude Opus 4.7	Anthropic	35th	1 bench · 1 src	—
154	Gemini 1.5 Pro	Google	35th	3 bench · 1 src	—
155	glm-4.7-flash	—	34th	3 bench · 1 src	$0.40
156	Llama 3.1-405B	Meta	34th	3 bench · 1 src	—
157	Qwen3-30B-A3B	Alibaba	33rd	3 bench · 1 src	$0.50
158	Mistral Large 2	Mistral	33rd	4 bench · 1 src	—
159	Qwen3.6 Plus	Alibaba	33rd	2 bench · 1 src	—
160	Phi-4	Microsoft	32nd	4 bench · 1 src	$0.50
161	Qwen2.5-72B	Alibaba	32nd	3 bench · 1 src	$0.408
162	GPT-5.4 nano	OpenAI	31st	5 bench · 1 src	—
163	gpt-oss-20b	OpenAI	31st	3 bench · 1 src	$0.30
164	GPT-4.1 nano	OpenAI	30th	6 bench · 2 src	$0.40
165	GPT-4 Turbo	OpenAI	30th	5 bench · 1 src	—
166	Claude 3 Opus	Anthropic	30th	6 bench · 2 src	$75.00
167	Qwen Plus	Alibaba	29th	3 bench · 1 src	—
168	Llama 3.3 70B	Meta	29th	3 bench · 1 src	free
169	Magistral Small 1.2	Mistral	28th	3 bench · 1 src	$1.50
170	qwen3-4b-instruct-2507	—	28th	3 bench · 1 src	—
171	Tülu 3 70B	Ai2	27th	3 bench · 1 src	—
172	Mistral Small 3.2	Mistral	27th	3 bench · 1 src	$0.18
173	Qwen3-4B	Alibaba	27th	2 bench · 1 src	$0.24
174	Gemma 3 27B	Google	26th	5 bench · 2 src	free
175	Qwen3 8B	Alibaba	26th	3 bench · 1 src	—
176	Llama 3.1-70B	Meta	25th	3 bench · 1 src	$0.72
177	Mistral Small 3	Mistral	25th	4 bench · 1 src	—
178	Qwen2.5-32B	Alibaba	24th	4 bench · 1 src	$0.90
179	DeepSeek-R1-Distill-Qwen-14B	DeepSeek	24th	4 bench · 1 src	$0.431
180	Mistral Small 3.1	Mistral	24th	4 bench · 1 src	—
181	Gemini 1.5 Flash	Google	23rd	4 bench · 1 src	—
182	glm-4.7-flash_none	—	23rd	3 bench · 1 src	—
183	WizardLM-2 8x22B	Microsoft	22nd	2 bench · 1 src	$0.493
184	GPT-4o mini	OpenAI	22nd	8 bench · 2 src	$0.60
185	GPT-4	OpenAI	21st	5 bench · 1 src	—
186	GPT-4 Turbo	OpenAI	21st	2 bench · 1 src	—
187	Qwen3.5-2B	Alibaba	21st	1 bench · 1 src	$0.16
188	phi-3-mini 3.8B	Microsoft	20th	1 bench · 1 src	—
189	Llama 2-7B	Meta	20th	1 bench · 1 src	$0.25
190	Llama 2-13B	Meta	19th	1 bench · 1 src	$0.50
191	Qwen Turbo	Alibaba	19th	3 bench · 1 src	$0.2006
192	Llama 3.2 90B	Meta	18th	3 bench · 1 src	$0.72
193	Qwen2-72B	Alibaba	18th	2 bench · 1 src	$0.90
194	Gemma 3 12B	Google	18th	3 bench · 1 src	$0.50
195	Claude 3 Sonnet	Anthropic	17th	3 bench · 1 src	$15.00
196	Llama 3-70B	Meta	17th	3 bench · 1 src	$2.75
197	Mistral Large	Mistral	16th	3 bench · 1 src	$6.001
198	Qwen2.5-7B	Alibaba	16th	3 bench · 1 src	$0.05
199	Claude 3.5 Haiku	Anthropic	15th	4 bench · 1 src	—
200	Qwen3-1.7B	Alibaba	15th	3 bench · 1 src	$0.10
201	DeepSeek-R1-Distill-Qwen-1.5B	DeepSeek	15th	3 bench · 1 src	free
202	Hermes 2 Theta Llama-3 70B	Nous Research	14th	3 bench · 1 src	—
203	Gemma 2 27B	Google	14th	3 bench · 1 src	$0.65
204	Claude 3 Haiku	Anthropic	13th	3 bench · 1 src	$1.25
205	GPT-4	OpenAI	13th	2 bench · 1 src	—
206	granite-4.0-micro	—	12th	3 bench · 1 src	—
207	GPT-3.5 Turbo	OpenAI	12th	6 bench · 1 src	$1.50
208	Claude 2	Anthropic	12th	4 bench · 2 src	—
209	Llama 3.1-8B	Meta	11th	4 bench · 1 src	$0.22
210	Mixtral 8x22B	Mistral	11th	2 bench · 1 src	$1.20
211	Gemini 1.0 Pro	Google	10th	3 bench · 1 src	—
212	Eurus-2-7B-PRIME	Tsinghua University	10th	1 bench · 1 src	—
213	Llama 3-8B	Meta	9th	4 bench · 1 src	$0.25
214	DeepSeek LLM 67B	DeepSeek	9th	4 bench · 1 src	—
215	Gemini 1.5 Flash 8B	Google	9th	2 bench · 1 src	—
216	Claude 2.1	Anthropic	8th	2 bench · 1 src	—
217	DBRX	Databricks	8th	2 bench · 1 src	$1.20
218	granite-4.0-1b	—	7th	3 bench · 1 src	—
219	Llama 3.2 1B	Meta	7th	3 bench · 1 src	$0.10
220	Gemma 3 4B	Google	6th	3 bench · 1 src	$0.2006
221	Yi-1.5-34B	01.AI	6th	2 bench · 1 src	—
222	Qwen1.5-32B	Alibaba	6th	1 bench · 1 src	—
223	Gemma 3 1B	Google	5th	3 bench · 1 src	—
224	Mixtral 8x7B	Mistral	5th	2 bench · 1 src	$0.50
225	Mistral NeMo	Mistral	4th	2 bench · 1 src	$0.17
226	Qwen1.5-72B	Alibaba	4th	1 bench · 1 src	$0.90
227	Mistral 7B	Mistral	3rd	4 bench · 1 src	$0.20
228	phi-3-medium 14B	Microsoft	3rd	2 bench · 1 src	—
229	Gemma 2 9B	Google	3rd	3 bench · 1 src	$0.20
230	DeepSeek-R1-0528	—	2nd	3 bench · 1 src	$0.20
231	Ministral 8B	Mistral	2nd	2 bench · 1 src	$0.15
232	Llama 2-70B	Meta	1st	3 bench · 1 src	$2.75
233	granite-4.0-350m	—	1st	3 bench · 1 src	—
234	Ministral 3B	Mistral	0th	2 bench · 1 src	$0.10
235	Yi-34B	01.AI	0th	2 bench · 1 src	$0.90

Common questions

What is the best ai for reasoning right now?: Ranked by the reasoning composite: GPQA Diamond, EBR-bench, Chess Puzzles and Mystery Game Puzzles.
How is this ranking computed?: It is the same family composite as the leaderboard. Each benchmark is weighted by how much of its spread survives its own measurement error.