Inference
概览
排行榜
基准测试
实验室
本地运行
Search
⌘K
中文
English
Русский
中文
12 Nov
2026
排行榜
/
智能体
最佳智能体 AI 模型
218 个模型 · 智能体指数。至少两项结果。
方法
编程
推理
智能体
多模态
数学
知识
指令遵循
低价
高速
免费
开放权重
长上下文
上榜模型
2026年11月
#
模型
智能体
指数
?
全部类别，满分 100。
输出 / 1M
?
每 1M 输出 token 的美元价格。
速度
?
每秒输出 token 数。
01
Gemini 4 Carbon
83.2
82.7
$
10.00
31
/s
02
Claude Fable 5.2
82.2
85.6
$
40.00
64
/s
03
GPT-6.1 Terra
78.6
77.1
$
5.00
46
/s
04
Muse Wonder Full
78.2
77.3
—
—
05
Claude Sonnet 5.5
77.5
77.2
$
10.00
107
/s
06
Gemini 4 Argon
77.3
78.5
$
10.00
28
/s
07
MA
Kimi K3.1 Pro
76.7
75.7
$
10.00
43
/s
08
GPT-6.1 Sol
76.0
78.0
$
10.00
34
/s
09
GPT-6 Astra
74.6
78.0
$
50.00
34
/s
10
TE
Hy4
74.5
72.0
$
1.25
52
/s
11
Claude Opus 5.5
74.1
78.6
$
20.00
76
/s
12
Qwen3.8 Max Preview
73.5
64.2
—
—
13
Muse Wonder
73.0
73.1
$
20.00
80
/s
14
DeepSeek V4.1 Flash 1017
72.2
68.5
$
0.40
261
/s
15
MiMo-V2.6-Pro
72.1
70.2
$
0.87
31
/s
16
GPT-5.6 Sol
71.9
73.0
$
22.00
64
/s
16
Grok 4.7
71.9
69.1
$
6.00
51
/s
18
MiniMax M3.1 Flash
71.8
69.3
$
0.60
100
/s
18
DeepSeek V4 Flash Vision Exp
71.8
60.4
$
1.32
74
/s
20
Claude Opus 5
71.6
73.8
$
25.00
48
/s
21
Muse Ember
71.4
64.8
$
0.50
158
/s
21
Claude Fable 5
71.4
74.0
$
50.00
56
/s
23
Claude Fable 5.1
71.1
76.2
$
50.00
69
/s
24
Muse Spark 1.3
70.4
70.1
$
4.25
76
/s
25
Claude Mythos 5
70.0
70.8
$
50.00
—
26
MiMo-V2.6-Flash
69.7
64.3
$
0.28
28
/s
26
GLM-5.3-Flash
69.7
63.1
$
0.078
122
/s
28
ST
Step 5 Preview
68.8
68.4
$
2.70
105
/s
29
GPT-6 Sol
68.6
72.8
$
10.00
61
/s
29
Claude Haiku 5.5
68.6
66.3
$
0.50
159
/s
31
Qwen3.8-Flash-Next
68.5
64.3
—
55
/s
32
Grok 4.6
68.4
67.8
$
6.00
56
/s
33
GLM-5.3
68.2
65.8
$
1.40
86
/s
33
GPT-5.6 Terra
68.2
68.9
$
12.00
66
/s
35
TM
Inkling 2 Small
67.6
66.4
$
1.00
—
36
GPT-6 Luna
66.9
65.5
$
0.50
66
/s
37
IN
Ling 3.1 Flash
66.6
67.2
$
0.135
81
/s
38
Claude Opus 4.8
66.5
67.7
$
25.00
69
/s
39
Qwen3.8 Max (0902)
66.4
63.4
$
6.00
33
/s
40
Mistral Large 4
66.3
60.7
$
2.10
113
/s
41
MA
Kimi K3
65.9
66.9
$
16.00
44
/s
42
GPT-5.5
65.8
69.1
$
30.00
52
/s
43
Gemini 3.8 Flash
65.3
67.8
$
4.13
68
/s
43
Claude Sonnet 5
65.3
64.2
$
10.00
75
/s
45
Gemini 3.7 Flash
65.1
68.5
$
3.75
85
/s
46
Muse Spark Open
64.8
65.1
$
4.25
86
/s
47
GPT-5.6 Luna
64.7
65.5
$
1.20
51
/s
47
Gemini 3.5 Flash
64.7
64.2
$
9.00
137
/s
49
Qwen3.8-Omni-Flash
64.5
62.7
$
0.47
41
/s
50
Qwen3.8 Max
63.7
65.6
—
41
/s
50
DeepSeek V4 Pro 0813
63.7
65.5
$
4.44
42
/s
52
AP
Apodex 1.1
63.5
58.9
—
—
52
Claude Opus 4.7 (Adaptive)
63.5
63.2
$
25.00
—
52
TE
Hy4 preview
63.5
66.4
$
2.50
67
/s
52
Claude Opus 4.6 (Adaptive)
63.5
60.0
—
—
56
DS
dots3-note Preview
63.3
63.4
Free
—
57
Muse Spark 1.2
62.7
64.9
$
4.25
123
/s
58
UN
Beam 1.0
62.3
58.8
—
—
59
GPT-5.4
62.1
64.8
$
15.00
38
/s
60
Muse Spark 1.1
61.9
64.6
$
4.25
152
/s
61
GLM-5.2
61.7
60.6
$
6.80
111
/s
62
Grok 4.5
61.6
63.7
$
6.00
53
/s
63
Claude Opus 4.7
60.6
62.0
$
25.00
47
/s
63
GPT-5.3 Codex
60.6
62.0
$
14.00
52
/s
65
Qwen3.7 Max
60.5
60.5
$
4.43
68
/s
66
Mistral Medium 3.5 128B
60.1
50.3
$
7.50
—
67
DeepSeek V4 Flash 0731
59.9
61.9
$
0.30
61
/s
67
AP
Apodex 1.1 Mini
59.9
55.7
Free
—
69
Claude Sonnet 4.6
59.7
57.0
$
15.00
44
/s
69
Gemini 3.6 Flash
59.7
63.7
$
3.75
121
/s
69
Claude Opus 4.5 Thinking
59.7
56.3
—
—
72
ST
Step 3.7 Flash
59.6
55.3
$
1.15
112
/s
73
MA
Kimi K2.6
59.5
58.2
$
3.41
81
/s
74
Qwen3.5 397B (Reasoning)
59.4
55.3
$
3.60
—
74
IN
Sage-A1.5
59.4
60.5
—
—
76
RA
Beam 1.0 501B-A23B
59.3
60.2
—
345
/s
77
Qwen 3.6 Max (preview)
59.2
57.2
$
6.16
75
/s
77
GLM-5.1
59.2
56.0
$
4.40
102
/s
79
Gemini 3.1 Pro
58.4
62.3
$
12.00
151
/s
79
MA
Kimi K2.7 Code
58.4
57.7
$
3.50
35
/s
81
MiMo-V2.5-Pro
58.2
56.3
$
1.80
47
/s
82
GPT-5.2-Codex
58.1
59.9
$
14.00
44
/s
83
Claude Opus 4.6
57.8
58.0
$
25.00
36
/s
84
GPT-5.1-Codex-Max
57.5
53.8
$
10.00
55
/s
85
Qwen3.8-27B
57.4
59.2
$
1.88
113
/s
85
MiniMax M3
57.4
55.7
$
1.20
79
/s
87
Gemini 3 Flash Preview
57.3
57.5
$
3.00
78
/s
87
Qwen3.7 Plus
57.3
57.7
$
1.28
51
/s
89
Claude Sonnet 4.5 Thinking
57.1
53.8
—
—
90
GLM-5-Turbo
57.0
51.9
$
4.00
25
/s
91
Qwen3.6-27B
56.7
52.1
$
3.20
60
/s
92
MiMo-V2.5
56.5
54.8
$
2.00
4
/s
93
LA
K-EXAONE 2.0
56.4
48.8
—
—
94
Gemini 3 Pro
56.2
57.9
$
12.00
109
/s
95
GPT-5.1-Codex
56.1
52.9
$
10.00
55
/s
95
GPT-5.4 mini
56.1
54.5
$
4.50
95
/s
97
DeepSeek V4 Pro 0423
55.9
57.6
$
1.78
89
/s
97
Grok Build 0.1
55.9
52.7
$
2.00
147
/s
97
Muse Spark
55.9
59.8
—
—
100
TM
Inkling-Small
55.7
55.1
$
1.20
80
/s
101
RA
Beam 1.0 501B-A23B FP8
55.4
59.2
—
—
102
GPT-5.2
55.3
59.0
$
14.00
45
/s
103
Qwen3.6 Plus
55.0
54.6
$
1.95
44
/s
103
RA
Beam 1.0 501B-A23B NVFP4
55.0
58.5
—
—
105
Gemini 3.5 Flash-Lite
54.7
52.0
$
2.50
130
/s
105
Claude Opus 4.5
54.7
52.0
$
25.00
39
/s
107
IN
Agents-A1-4B
54.1
46.8
—
—
108
PM
Ternary Bonsai 2 27B
54.0
52.2
$
0.50
21
/s
109
IN
Ling 3.0 Flash
53.5
50.9
$
0.063
73
/s
110
Qwen3.5 397B
53.3
53.8
$
3.60
59
/s
110
MiMo-V2-Pro
53.3
51.8
—
—
110
Muse Glimmer 30B
53.3
53.0
$
1.20
90
/s
113
GPT-5
53.1
55.0
$
10.00
76
/s
114
Qwen3.6-35B-A3B
52.7
50.6
$
0.90
65
/s
115
TE
Hy3 Preview
52.6
52.3
$
0.60
69
/s
116
Gemini 3.1 Pro Preview
52.5
61.5
$
12.00
77
/s
117
GPT-5.1
52.4
55.3
$
10.00
50
/s
118
MiniMax M2.7
52.0
51.2
$
0.84
35
/s
119
GLM-5
51.6
51.6
$
3.20
45
/s
120
Qwen3.5-122B-A10B
51.4
51.9
$
2.08
70
/s
121
TM
Inkling
51.2
54.8
$
4.05
85
/s
122
Grok 4.20
50.7
52.8
$
2.50
72
/s
123
Grok 4.3
50.6
53.3
$
2.50
61
/s
123
MA
Kimi K2.5 (Reasoning)
50.6
53.0
$
3.00
—
125
GPT-5.4 nano
50.5
50.9
$
1.25
91
/s
125
GLM-5V-Turbo
50.5
48.1
$
4.00
30
/s
127
Qwen3.5-27B
50.3
51.3
$
2.60
69
/s
128
Claude Haiku 4.5 Thinking
50.2
43.5
—
—
129
UP
Solar Pro 4
49.9
52.1
$
0.60
44
/s
130
GPT-5 (high)
49.7
53.6
$
10.00
76
/s
131
RI
Ripple-65B Preview
49.5
52.3
$
0.60
63
/s
132
MA
Kimi K2.5
49.4
52.2
$
3.32
36
/s
133
Claude Sonnet 4.5
49.3
47.9
$
15.00
45
/s
133
IN
Mercury 2.5
49.3
46.2
$
0.15
125
/s
135
Grok 4.20 (Reasoning)
48.9
55.5
—
—
135
Nemotron 3 Ultra
48.9
50.8
Free
—
135
Claude Haiku 4.5
48.9
43.2
$
5.00
88
/s
138
SP
Spark-X2.6-4B
48.8
48.7
—
—
138
GLM-4.7
48.8
50.0
$
2.20
28
/s
138
GLM-4.6
48.8
41.8
$
2.00
30
/s
141
Gemini 3 Flash
48.5
55.0
$
3.00
198
/s
141
MiMo-V2-Flash
48.5
43.4
—
129
/s
141
MiniMax M2.1
48.5
49.1
$
1.20
34
/s
144
Grok 4.1 Fast (Reasoning)
48.4
47.3
—
—
145
AA
Trinity-Large-Preview
48.3
42.2
$
1.00
—
145
MA
Kimi K2.5 Thinking
48.3
52.6
—
—
147
MA
Kimi K2 Thinking
48.0
48.2
$
2.50
48
/s
147
Grok 4
48.0
50.0
—
78
/s
149
UP
Solar Open 2
47.9
53.3
—
—
150
Nemotron 3.5 Super VL 120B A12B
47.8
51.0
—
—
151
DeepSeek V3.2 (Thinking)
47.7
49.4
$
2.19
—
152
Qwen3.5-35B-A3B
47.3
49.4
$
1.00
43
/s
153
Mistral Medium 3.5
47.1
42.7
$
7.50
92
/s
153
Qwen3 Max
47.1
45.0
$
3.90
62
/s
155
UP
Solar Pro 3
47.0
39.2
$
0.60
114
/s
156
IN
Ling 2.6 Flash
46.9
37.0
—
104
/s
156
MiMo-V2-Omni
46.9
47.2
—
—
158
LI
LFM3-1.7B
46.4
41.5
$
0.20
218
/s
159
ST
A.X K2
46.3
51.9
—
—
159
LI
LFM2.5-8B-A1B
46.3
30.6
—
335
/s
161
Nemotron 3.5 Super VL 120B A12B NVFP4
45.1
50.8
—
—
162
Gemma 4 31B
44.9
47.9
Free
35
/s
163
MA
Kimi K2 Instruct
44.6
43.2
—
—
164
IN
Mercury 3
44.4
52.3
$
0.35
295
/s
165
AA
Trinity-Large-Thinking
44.0
43.5
$
0.80
138
/s
166
Claude 4 Sonnet
43.9
39.5
$
15.00
40
/s
167
DeepSeek V3.2
43.1
44.2
$
0.38
19
/s
167
LA
K-Exaone
43.1
42.3
—
—
169
Grok 4 Fast (Reasoning)
42.9
45.3
—
—
169
Grok Code Fast 1
42.9
41.2
$
1.50
172
/s
171
Nemotron 3.5 Bolt 74B-A9B
42.3
46.4
—
—
172
IB
Granite 4.2 30B
42.0
41.6
—
—
173
CO
Command A+
41.7
40.2
$
1.50
161
/s
174
GPT-5 mini
41.4
48.4
$
2.00
78
/s
175
Gemini 3.1 Flash Lite Preview
41.3
49.6
$
1.50
151
/s
176
Grok 4.1 Fast
41.1
33.3
$
0.50
138
/s
177
Gemini 2.5 Pro
40.4
46.6
$
10.00
91
/s
178
GPT-OSS 120B
40.1
43.4
$
0.60
34
/s
179
GPT-OSS 20B
39.8
41.7
$
0.13
310
/s
180
ME
LongCat-Flash-Lite-Sparse
39.6
39.3
—
—
181
OP
MiniCPM5-2B
38.2
37.1
—
—
182
DeepSeek V3.1
38.0
37.9
—
98
/s
183
GPT-4.1
37.8
38.0
$
8.00
74
/s
184
Llama 4 Maverick
36.9
34.4
$
1.00
61
/s
185
IB
Granite 4.2 8B
36.8
37.6
$
0.15
101
/s
186
SA
Sarvam 105B
36.4
36.7
—
191
/s
187
Grok 4.1 Fast Non Reasoning
35.5
36.2
—
—
188
GPT-4.1 mini
35.3
34.5
$
1.60
72
/s
189
IN
LLaDA2.2-mini
34.8
24.3
—
—
190
DeepSeek V3 0324
34.0
37.1
—
—
190
Gemma 4 26B A4B
34.0
44.5
Free
—
192
Mistral Small 4
33.6
37.9
$
0.60
114
/s
193
Nemotron 3 Nano Omni 30B A3B
33.4
38.6
—
228
/s
194
IB
Granite 4.2 3B
33.3
33.2
—
—
195
Nemotron 3.5 Lightning 30B A3B
33.0
28.1
—
—
196
LI
LFM2.5-2.6B
32.0
29.3
$
0.50
189
/s
196
LI
LFM2.5-24B-A2B
32.0
39.8
$
0.70
280
/s
198
CO
North Mini Code
30.8
38.0
Free
70
/s
199
Magistral Medium
30.6
33.7
—
—
200
Gemma 4 12B
30.5
39.0
—
—
201
Nemotron 3.5 Lightning 30B A3B NVFP4
30.2
37.0
—
—
202
Mistral Large 3
29.7
34.3
$
1.50
61
/s
203
Llama4 Maverick Instruct Basic
29.4
36.0
—
—
204
DeepSeek-R1
27.6
38.9
$
2.50
18
/s
205
Llama 3.3 70B
27.2
28.2
—
—
206
Nemotron 3 Nano 30B
26.8
32.9
—
270
/s
206
SA
Sarvam 30B
26.8
30.3
—
—
208
GPT-4o
26.7
28.7
$
10.00
66
/s
208
Mistral Small 3.1
26.7
24.2
—
28
/s
210
Claude 3.5 Haiku
26.6
24.1
—
—
211
DeepSeek V3
26.0
29.9
$
1.10
37
/s
212
CO
Command A
25.3
23.9
—
—
212
Gemma 4 E4B
25.3
28.2
—
—
212
Gemma 4 E2B
25.3
21.9
—
—
215
Magistral Small
24.8
31.4
—
—
216
GPT-4.1 nano
24.2
26.3
$
0.40
56
/s
217
Llama 4 Scout
23.6
26.8
$
0.30
63
/s
218
Gemma 3 27B
21.9
25.3
—
31
/s
Show all 218
