AI model rankings

Which AI model is the most capable? Rankings built on independent tests run by Epoch AI, side by side with prices so you can find the best model for your budget.

Quick picks

Category
#ModelOverall indexiGPQA DiamondiSWE-bench VerifiediInputiOutputiReleased
1
166.6
95.8%—$10$50Sep 2026
2
165.0
——$10$50Sep 2026
3
163.6
85.9%—$10$50Jun 2026
4
Claude Opus 5Anthropic
162.7
93.9%—$5$25Jul 2026
5
162.5
——$30$180Apr 2026
6
162.0
93.5%—$2$10Jul 2026
7
159.3
93.3%—$2$12Jul 2026
8
GPT-5.5OpenAI
159.3
90.7%—$5$30Apr 2026
9
159.1
94.6%—$30$180Mar 2026
10
158.3
91.0%—$5$25May 2026
11
Gemini 3.7 FlashGoogle DeepMind
157.7
94.8%—$0.75$3.75Aug 2026
12
Kimi K3Moonshot AI
157.7
93.1%—$3$15Jul 2026
13
Gemini 3.8 FlashGoogle DeepMind
157.1
95.4%—$0.75$3.75Sep 2026
14
156.9
——$1.25$4.25Sep 2026
15
GPT-5.4OpenAI
156.9
93.3%76.9%$2.50$15Mar 2026
16
156.8
—74.8%$1.75$14Feb 2026
17
Qwen3.8 Max (0902)Alibaba Qwen
156.7
92.7%—$2$6Sep 2026
18
156.5
94.0%—$2$6Aug 2026
19
156.4
90.2%83.5%$5$25Apr 2026
20
156.3
91.6%—$0.20$1.20Jul 2026
21
156.3
90.5%—$2$10Jun 2026
22
GLM 5.3Z.ai (Zhipu)
155.6
90.9%—$1.40$4.40Aug 2026
23
155.4
91.7%77.6%$0.66$1.98Aug 2026
24
155.4
90.5%78.7%$5$25Feb 2026
25
155.4
——$21$168Dec 2025
26
155.2
——$1.25$4.25Aug 2026
27
155.0
——$0.30$1.20Sep 2026
28
Gemini 3.1 Pro PreviewGoogle DeepMind
154.9
94.4%—$2$12Feb 2026
29
Gemini 3.5 FlashGoogle DeepMind
154.6
92.8%79.3%$1.50$9May 2026
30
154.5
91.0%—$0.018$0.32Jul 2026
31
Gemini 3.6 FlashGoogle DeepMind
154.4
94.1%—$0.75$3.75Jul 2026
32
154.3
——$1.25$4.25Jul 2026
33
154.0
93.4%—$2$6Jul 2026
34
Qwen3.7 MaxAlibaba Qwen
153.7
90.9%77.3%$1.48$4.43May 2026
35
153.5
91.4%73.8%$1.75$14Dec 2025
36
152.3
87.4%75.2%$3$15Feb 2026
37
152.0
——$1.25$2.50Mar 2026
38
GLM 5.3 FlashZ.ai (Zhipu)
151.9
90.2%—$0.15$0.50Aug 2026
39
GLM 5.2Z.ai (Zhipu)
151.8
91.9%78.7%$0.14$3.99Jun 2026
40
Gemini 3 Flash PreviewGoogle DeepMind
151.8
89.4%75.4%$0.50$3Dec 2025
41
Kimi K2.6Moonshot AI
151.0
90.8%76.7%$0.65$3.41Apr 2026
42
GPT-5 ProOpenAI
150.3
——$15$120Oct 2025
43
Inkling SmallThinking Machines Lab
150.1
88.5%—$0.45$1.20Jul 2026
44
150.1
86.0%76.7%$5$25Nov 2025
45
Kimi K2.7 CodeMoonshot AI
150.0
87.9%—$0.66$3.30Jun 2026
46
GPT-5OpenAI
150.0
86.2%73.6%$1.25$10Aug 2025
47
GLM 5.1Z.ai (Zhipu)
149.9
89.9%74.2%$1.40$4.40Apr 2026
48
GPT-5.1OpenAI
149.7
87.6%68.0%$1.25$10Nov 2025
49
Qwen3.8 27BAlibaba Qwen
149.4
——$0.42$3Aug 2026
50
149.3
87.4%76.7%$1.03$6.16Apr 2026
51
148.8
86.9%—$0.75$4.50Mar 2026
52
InklingThinking Machines Lab
148.6
88.3%—$1$4.05Jul 2026
53
Kimi K2.5Moonshot AI
148.1
—73.8%$0.45$2.25Jan 2026
54
Qwen3.6 PlusAlibaba Qwen
147.7
88.4%57.9%$0.33$1.95Apr 2026
55
Qwen3.7 PlusAlibaba Qwen
147.4
87.9%—$0.32$1.28Jun 2026
56
o3 ProOpenAI
147.4
——$20$80Jun 2025
57
MiniMax M3MiniMax
147.0
90.9%—$0.30$1.20May 2026
58
o3OpenAI
146.9
81.8%62.3%$2$8Apr 2025
59
146.8
84.8%—$0.30$1.80Apr 2026
60
146.8
82.3%71.3%$3$15Sep 2025
61
146.7
——$0.27$1.08Feb 2026
62
Qwen3.5 397B A17BAlibaba Qwen
146.6
86.4%—$0.55$3.50Feb 2026
63
Qwen3.6 27BAlibaba Qwen
146.5
85.9%—$0.32$3.20Apr 2026
64
146.3
——$0.28$0.42Dec 2025
65
146.2
85.4%—$0.60$2.40Jun 2026
66
Kimi K2 ThinkingMoonshot AI
146.0
——$0.60$2.50Nov 2025
67
145.8
——$0.21$0.84Mar 2026
68
145.8
78.5%—$0.20$1.25Mar 2026
69
GLM 5Z.ai (Zhipu)
145.8
87.8%72.1%$0.60$1.92Feb 2026
70
o4 MiniOpenAI
145.6
79.6%—$1.10$4.40Apr 2025
71
145.5
75.0%64.7%$0.25$2Aug 2025
72
Gemini 2.5 ProGoogle DeepMind
145.3
85.3%57.6%$1.25$10Jun 2025
73
Gemini 3.5 Flash LiteGoogle DeepMind
145.1
83.3%—$0.30$2.50Jul 2026
74
145.0
——$0.27$0.41Sep 2025
75
Qwen3.7 FlashAlibaba Qwen
144.6
82.3%—$0.03$0.13Jul 2026
76
Gemini 3.1 Flash LiteGoogle DeepMind
144.5
81.8%—$0.25$1.50May 2026
77
144.1
77.3%73.3%$15$75Aug 2025
78
Qwen3.5-FlashAlibaba Qwen
144.0
82.3%—$0.065$0.26Feb 2026
79
Qwen3.6 35B A3BAlibaba Qwen
143.9
84.8%—$0.15$1Apr 2026
80
143.9
80.1%—$0.23$2.30Jul 2025
81
GLM 4.7Z.ai (Zhipu)
143.5
83.3%—$0.60$2.20Dec 2025
82
Qwen3.6 FlashAlibaba Qwen
143.3
83.3%—$0.19$1.13Apr 2026
83
Gemini 2.5 FlashGoogle DeepMind
143.0
——$0.30$2.50Jun 2025
84
Gemma 4 31BGoogle DeepMind
142.7
75.8%—$0.09$0.34Apr 2026
85
Qwen3.5-35B-A3BAlibaba Qwen
142.5
83.5%—$0.16$1.30Feb 2026
86
142.4
71.2%—$1$5Oct 2025
87
Qwen3 MaxAlibaba Qwen
142.4
72.6%—$0.78$3.90Sep 2025
88
Gemma 4 26B A4B Google DeepMind
141.9
73.2%—$0.077$0.26Apr 2026
89
o1OpenAI
141.9
76.8%—$15$60Dec 2024
90
141.7
79.2%—$3$15May 2025
91
141.4
——$1.50$7.50Apr 2026
92
R1 0528DeepSeek
141.3
76.3%—$0.50$2.15May 2025
93
o3 MiniOpenAI
140.4
77.0%—$1.10$4.40Jan 2025
94
Kimi K2 0905Moonshot AI
140.1
——$0.60$2.50Sep 2025
95
139.9
——$0.25$0.95Aug 2025
96
139.9
75.8%—$0.037$0.17Aug 2025
97
139.6
70.1%—$0.20$2.40Aug 2025
98
Qwen3.5-9BAlibaba Qwen
139.5
79.0%—$0.10$0.15Mar 2026
99
139.4
69.4%—$0.05$0.40Aug 2025
100
139.4
70.7%—$0.088$0.35Jul 2025
101
Qwen3 32BAlibaba Qwen
138.5
65.7%—$0.08$0.28Apr 2025
102
Qwen3 14BAlibaba Qwen
138.2
63.8%—$0.12$0.24Apr 2025
103
137.8
60.8%—$0.018$0.09Aug 2025
104
137.4
61.7%—$0.048$0.19Jul 2025
105
GPT-4.1OpenAI
136.8
66.9%48.5%$2$8Apr 2025
106
Qwen3 8BAlibaba Qwen
136.2
56.8%—$0.12$0.46Apr 2025
107
DeepSeek V3DeepSeek
136.0
71.2%—$0.26$1.03Dec 2024
108
135.0
65.8%—$0.40$1.60Apr 2025
109
134.1
——$0.40$2May 2025
110
Gemini 2.5 Flash LiteGoogle DeepMind
133.9
——$0.10$0.40Jul 2025
111
132.2
——$0.19$0.65Apr 2025
112
Phi 4Microsoft
130.4
56.1%—$0.07$0.14Jan 2025
113
Gemma 3 27BGoogle DeepMind
130.0
47.7%—$0.08$0.45Mar 2025
114
129.6
48.9%—$0.10$0.40Apr 2025
115
129.6
——$0.10$0.30Apr 2025
116
129.0
49.2%31.0%$2.50$10Nov 2024
117
129.0
49.1%—$0.36$0.40Sep 2024
118
127.3
47.4%—$0.10$0.32Dec 2024
119
127.3
46.6%—$10$30Apr 2024
120
Mistral Small 3Mistral AI
127.1
47.3%—$0.05$0.08Jan 2025
121
126.6
37.7%—$0.15$0.60Jul 2024
122
125.9
44.2%—$0.40$0.40Jul 2024
123
GPT-4OpenAI
125.9
35.7%—$30$60May 2023
124
Gemma 3 12BGoogle DeepMind
123.5
39.5%—$0.05$0.15Mar 2025
125
Gemma 2 27BGoogle DeepMind
122.1
36.5%—$0.65$0.65Jul 2024
126
122.0
51.3%—$0.50$1.50Dec 2025
127
122.0
——$2$6Apr 2024
128
119.4
——$0.66$1Nov 2024
129
119.3
——$0.15$0.60Aug 2024
130
Mistral NemoMistral AI
118.6
——$0.019$0.03Jul 2024
131
118.5
28.0%—$1$2Jan 2024
132
118.4
35.5%—$0.10$0.20Oct 2024
133
118.1
25.3%—$0.10$0.10Dec 2025
134
116.5
27.0%—$0.05$0.08Jul 2024
135
Gemma 3 4BGoogle DeepMind
116.0
23.2%—$0.05$0.10Mar 2025
136
102.0
23.9%—$0.027$0.20Sep 2024

per 1M tokens · USD

How to read these rankings

Scores come from tests run independently by Epoch AI rather than figures published by the AI companies themselves. When a model was tested with several settings (for example different reasoning efforts), we show its best result. Very new models appear once they have been tested, usually within a few weeks of release.

Overall index
Epoch Capabilities Index: one score that combines dozens of benchmarks, so models tested on different tests can still be compared. Higher is better; it is an index, not a percentage.
SWE-bench Verified
Can the model fix real bugs in real open-source software projects? Share of tasks solved.
GPQA Diamond
Very hard multiple-choice questions in biology, physics and chemistry, written by PhD experts. Random guessing scores 25%.
AIME math
Competition-level math problems in the style of the American Invitational Mathematics Examination. Share solved.
FrontierMath
Original, unpublished math problems written by professional mathematicians, from advanced to research level. Share solved.
SimpleQA Verified
Short factual questions. Measures how often the model gets facts right instead of making something up.

Benchmark data: Epoch AI, licensed CC BY 4.0 · Last update: Sep 30, 2026 · How we work

CodingMathScienceCheapFreeOpen sourceLong contextImages & PDFs

Best AI for your language →