M
ModelSage
排行榜/代码能力

AI代码能力排行榜

基于AA Coding Index编程评分排名，找到最适合开发的大模型。

分数怎么算：分数 = 该模型的 AA Coding Index ÷ 当前最高 Index × 100，榜首正好 100 分。

数据更新于 2026-10-14，每天北京时间凌晨自动刷新

综合能力代码能力Agent 智能体Arena 人类偏好视觉/多模态搜索能力文档理解响应速度性价比
RankModelScore
1
Claude Fable 5.1 (max with fallback)
100.0
2
Claude Fable 5.1 (xhigh with fallback)
98.8
3
Claude Fable 5.1 (high with fallback)
97.0
4
GPT-5.6 Sol (xhigh)
96.1
5
GPT-6 Astra (high)
95.8
6
Claude Opus 5 (max)
95.5
7
GPT-6 Astra (xhigh)
95.2
8
GPT-6 Astra (max)
94.9
9
GPT-6 Astra (medium)
94.9
10
GPT-5.6 Sol (max)
94.9
11
Grok 4.6 (high)
94.9
12
Claude Fable 5.1 (medium with fallback)
94.6
13
GPT-5.6 Sol (high)
94.5
14
Gemini 3.8 Flash (high)
94.4
15
Claude Opus 5 (xhigh)
94.3
16
GPT-5.6 Terra (max)
94.0
17
Claude Opus 5 (high)
93.9
18
GPT-6 Astra (low)
93.7
19
Claude Fable 5 (with fallback)
93.6
20
Qwen3.8 Max (0902)
93.5
21
GPT-5.6 Sol (medium)
93.5
22
Kimi K3 (max)
93.4
23
Grok 4.6 (xhigh)
93.3
24
Gemini 3.7 Flash (high)
93.1
25
Claude Fable 5.1 (low with fallback)
92.3
26
GLM-5.3 (max)
92.2
27
GPT-5.5 (xhigh)
91.3
28
Grok 4.6 (medium)
91.3
29
Claude Opus 5 (medium)
91.1
30
Claude Opus 4.8 (max)
91.0
31
Qwen3.8-Flash-Next
90.8
32
Gemini 3.8 Flash (medium)
90.7
33
Gemini 3.8 Flash (low)
90.0
34
GLM-5.3-Flash
89.6
35
Grok 4.5 (high)
88.8
36
Qwen3.8 2.4T A95B
88.7
37
Kimi K3 (low)
88.5
38
Gemini 3.7 Flash (medium)
88.0
39
Qwen3.8 Max
87.8
40
GPT-5.5 (medium)
87.7
41
Claude Sonnet 5 (max)
87.6
42
GPT-5.6 Luna (max)
87.6
43
Gemini 3.7 Flash (low)
87.5
44
GPT-5.5 (high)
87.4
45
GPT-5.4 (xhigh)
86.9
46
GPT-5.6 Terra (xhigh)
86.5
47
Gemini 3.5 Flash (high)
85.9
48
GPT-5.6 Sol (low)
85.4
49
DeepSeek V4 Pro 0813 (max)
84.8
50
Gemini 3.6 Flash (high)
84.8
51
DeepSeek V4 Flash 0731 (max)
84.5
52
GLM-5.2 (max)
84.2
53
Qwen3.8 27B (xhigh)
84.0
54
GPT-5.6 Luna (xhigh)
83.9
55
Gemini 3.1 Pro Preview
83.8
56
GPT-5.6 Terra (high)
82.9
57
Claude Opus 4.7 (non-reasoning, high)
82.8
58
Claude Opus 5 (low)
82.2
59
Grok 4.6 (low)
81.4
60
Qwen3.7 Max
80.8
61
GPT-5.3 Codex (xhigh)
80.8
62
DeepSeek V4 Flash Vision (max)
80.5
63
GPT-5.5 (low)
79.8
64
GPT-5.6 Sol (non-reasoning)
79.8
65
GPT-5.6 Terra (medium)
79.3
66
Claude Opus 4.7 (max)
78.8
67
Claude Sonnet 4.6 (max)
78.3
68
GPT-5.6 Luna (high)
77.6
69
GPT-5.5 (non-reasoning)
75.8
70
Claude Opus 4.6 (non-reasoning, high)
74.8
71
Kimi K2.6
74.5
72
Kimi K2.7 Code
74.3
73
MiMo-V2.5-Pro
73.5
74
Claude Opus 4.5
72.8
75
GPT-5.2 (xhigh)
72.8
76
DeepSeek V4 Pro (max)
72.7
77
MiniMax-M3
72.2
78
Hy3
72.1
79
Claude Opus 4.6 (max)
71.7
80
Claude Sonnet 4.6 (non-reasoning, high)
71.7
81
Gemini 3.5 Flash (minimal)
71.7
82
Qwen3.8 27B (low)
71.2
83
GPT-5.6 Terra (low)
71.1
84
GPT-5.1 (high)
70.7
85
MiMo-V2.5
69.7
86
GPT-5.4 mini (xhigh)
69.2
87
Qwen3.8 27B (medium)
69.0
88
Qwen3.6 Max Preview
68.7
89
Qwen3.6 Plus
68.7
90
DeepSeek V4 Flash (max)
68.7
91
GPT-5.4 nano (xhigh)
68.5
92
GLM-5.1
68.5
93
Qwen3.7 Plus
68.4
94
GPT-5.4 (low)
67.7
95
GPT-5.2 (medium)
67.7
96
GLM-5
67.7
97
Qwen3.6 27B
66.8
98
Claude Sonnet 4.6 (non-reasoning, low)
66.7
99
GPT-5.5 Instant (May 2026)
66.7
100
DeepSeek V4 Pro (high)
65.7
101
Gemini 3 Pro Preview (high)
65.7
102
MiniMax-M2.7
65.3
103
Claude Opus 4.5 (non-reasoning)
64.7
104
Grok 4.20 0309
64.7
105
GPT-5.6 Terra (non-reasoning)
64.0
106
Claude 4.5 Sonnet
63.9
107
DeepSeek V4 Flash (high)
62.7
108
Gemini 3.5 Flash (medium)
62.7
109
GLM-5 (non-reasoning)
62.7
110
GPT-5.6 Luna (medium)
62.2
111
Gemini 3 Flash
61.7
112
Kimi K2.6 (non-reasoning)
60.7
113
GPT-5 Codex (high)
60.7
114
GPT-5 (medium)
60.7
115
GPT-5.4 (non-reasoning)
60.7
116
Grok 4.20 0309 v2
60.7
117
Grok 4
60.7
118
Gemini 3.5 Flash-Lite
60.3
119
Qwen3.5 397B A17B
59.9
120
GPT-5.2 Codex (xhigh)
59.7
121
o3
59.7
122
DeepSeek V4 Pro (non-reasoning)
58.6
123
Qwen3.5 397B A17B (non-reasoning)
57.6
124
DeepSeek V3.2
57.6
125
MiMo-V2.5-Pro (non-reasoning)
57.6
126
GLM-5.1 (non-reasoning)
57.6
127
Mistral Medium 3.5
57.4
128
MiniMax-M2.5
56.6
129
Kimi K2.5
56.6
130
GPT-5.1 Codex (high)
56.6
131
Claude 4.1 Opus
55.9
132
DeepSeek V4 Flash (non-reasoning)
55.6
133
Gemini 3 Pro Preview (low)
55.6
134
GPT-5.4 mini (medium)
55.6
135
Hy3-preview
55.6
136
Qwen3.5 122B A10B
54.8
137
Qwen3.8 27B (non-reasoning)
54.6
138
GPT-5.1 Codex mini (high)
54.6
139
GPT-5.4 nano (medium)
54.6
140
GPT-5.6 Luna (low)
53.9
141
Gemma 4 31BFREE
53.7
142
Qwen3.5 27B
53.6
143
DeepSeek V3.2 (non-reasoning)
53.6
144
GPT-5 (high)
53.6
145
GLM 5V Turbo
53.6
146
Claude 4.5 Haiku
53.0
147
Qwen3.5 27B (non-reasoning)
52.6
148
DeepSeek V3.1 Terminus (non-reasoning)
52.6
149
Gemini 3 Flash (non-reasoning)
52.6
150
GPT-5.2 (non-reasoning)
52.6
151
Hy3-preview (non-reasoning)
52.6
152
GLM-4.7
52.6
153
DeepSeek V3.1 Terminus
51.9
154
Grok 4.3
51.7
155
Claude 4 Opus
51.6
156
DeepSeek V3.2 Exp
51.6
157
Kimi K2 Thinking
51.6
158
Qwen3.6 35B A3B
51.4
159
Gemma 4 31B (non-reasoning)
50.6
160
Grok 4.3 (medium)
50.6
161
GLM-4.7 (non-reasoning)
50.6
162
Qwen3.5 122B A10B (non-reasoning)
49.6
163
GPT-5.5 Instant (June 2026)
49.4
164
Claude 4.5 Sonnet (non-reasoning)
48.6
165
MiniMax-M2.1
48.6
166
GPT-5 mini (medium)
48.6
167
GLM-4.6 (non-reasoning)
48.6
168
GPT-5.6 Luna (non-reasoning)
48.0
169
Gemma 4 26B A4B
47.8
170
MiMo-V2-Flash
47.6
171
Claude 4.5 Haiku (non-reasoning)
46.6
172
Qwen3 Coder Next
45.7
173
Qwen3.5 35B A3B
45.5
174
GPT-5 (low)
45.5
175
Doubao Seed Code
45.5
176
Grok 4.3 (low)
45.5
177
Qwen3.6 35B A3B (non-reasoning)
44.5
178
MiniMax-M2
44.5
179
Grok 4.3 (non-reasoning)
43.6
180
DeepSeek V3.2 Exp (non-reasoning)
43.5
181
DeepSeek V3.1
43.5
182
Gemma 4 26B A4B (non-reasoning)
43.5
183
GLM-4.6
43.5
184
Gemini 3.1 Flash-Lite
42.7
185
Qwen3 Max Thinking
42.5
186
DeepSeek V3.1 (non-reasoning)
42.5
187
GPT-5.4 nano (non-reasoning)
42.5
188
Gemini 2.5 Pro
41.7
189
Kimi K2 0905
41.5
190
GPT-5.1 (non-reasoning)
40.5
191
Grok 4.20 0309 (non-reasoning)
39.5
192
GLM-4.7-FlashFREE
39.5
193
GLM-4.5
39.5
194
Qwen3.5 Omni Plus
38.5
195
Qwen3.6 27B (non-reasoning)
38.5
196
Claude 3.7 Sonnet (non-reasoning)
38.5
197
Qwen3 Max
37.5
198
GLM-4.5-Air
37.5
199
Qwen3 Max (Preview)
36.5
200
Qwen3.5 9B
35.7
201
Qwen3 Coder 480B
35.5
202
Kimi K2.5 (non-reasoning)
35.5
203
Grok 4 Fast
35.5
204
gpt-oss-120b (high)
35.0
205
Qwen3.5 9B (non-reasoning)
34.5
206
Qwen3.5 4B
34.5
207
Gemma 4 12B
34.5
208
GPT-5 (minimal)
34.5
209
GPT-5.4 mini (non-reasoning)
34.5
210
Qwen3 Max Thinking (Preview)
33.5
211
GPT-5 nano (medium)
33.5
212
Grok 3 mini Reasoning (high)
33.5
213
Mistral Small 4
32.6
214
Mistral Small 3.1
32.5
215
Grok 4.20 0309 v2 (non-reasoning)
32.4
216
DeepSeek R1 0528
31.4
217
Kimi K2
31.4
218
DeepSeek R1 (Jan)
30.8
219
Qwen3 235B 2507
30.4
220
Qwen3 Coder 30B A3B
30.4
221
o4-mini (high)
30.4
222
GPT-5 mini (minimal)
29.4
223
GLM-4.6V
29.4
224
Gemini 2.5 Flash
28.4
225
GPT-4.1
28.4
226
DeepSeek V3 (Dec)
28.0
227
Gemini 2.5 Flash-Lite (Sep)
27.4
228
o1
27.4
229
DeepSeek V3 0324
26.8
230
gpt-oss-20b (high)
26.8
231
Gemini 2.5 Flash (non-reasoning)
26.4
232
GPT-5 nano (high)
26.4
233
Grok 4 Fast (non-reasoning)
26.4
234
Qwen3 235B A22B 2507
26.2
235
Qwen3 VL 235B A22B (推理)
25.4
236
Qwen3.5 4B (non-reasoning)
25.4
237
Gemma 4 12B (non-reasoning)
25.4
238
Grok 3
25.4
239
Qwen3.5 35B A3B (non-reasoning)
24.4
240
Mistral Small 4 (non-reasoning)
24.4
241
Mistral Large 3
24.3
242
Qwen3 Next 80B A3B (推理)
23.4
243
Qwen3.5 Omni Flash
21.4
244
Qwen3 VL 32B (指令)
21.4
245
GPT-4o (Nov)
21.4
246
GPT-4o (Aug)
21.4
247
o3-mini (high)
20.6
248
Qwen3 VL 32B (推理)
20.4
249
Qwen3 Next 80B A3B (指令)
20.4
250
Gemini 2.5 Flash-Lite (Sep) (non-reasoning)
20.4
251
Gemma 4 E4B (non-reasoning)
20.4
252
GPT-4.1 mini
20.4
253
Qwen3 VL 235B A22B (指令)
19.4
254
Qwen3 30B (non-reasoning)
19.4
255
o3-mini
19.4
256
GPT-5 nano (minimal)
19.4
257
Seed-OSS-36B-Instruct
19.4
258
GLM-4.5V (non-reasoning)
19.4
259
Qwen3 32B
18.6
260
GPT-5 mini (high)
18.5
261
Qwen3 235B
18.3
262
Qwen3 235B (non-reasoning)
18.3
263
Qwen3 VL 30B A3B (指令)
18.3
264
Qwen3 30B A3B 2507 (non-reasoning)
18.3
265
ERNIE 4.5 300B A47B
18.3
266
Ministral 3 14B
17.4
267
Qwen3 VL 30B A3B (推理)
17.3
268
Qwen3 14B (non-reasoning)
17.3
269
gpt-oss-120b (low)
17.3
270
GLM-4.5V
17.3
271
Qwen2.5 72B
16.3
272
Gemini 2.5 Flash-Lite
16.3
273
Magistral Small 1.2
16.3
274
gpt-oss-20b (low)
16.3
275
Qwen3 14B
16.1
276
Mistral Small 3.2
15.9
277
Qwen3 VL 8B (推理)
15.3
278
Qwen3 Omni 30B A3B (推理)
15.3
279
Qwen3.5 2B
15.3
280
Qwen3.5 2B (non-reasoning)
15.3
281
Mistral Medium 3
15.3
282
GPT-4.1 nano
15.3
283
GLM-4.7-Flash (non-reasoning)FREE
15.3
284
Qwen3 30B A3B 2507
14.3
285
MiniMax M1 80k
14.3
286
GLM-4.6V (non-reasoning)
14.3
287
Qwen3 30B
13.3
288
Qwen3 VL 8B (指令)
13.3
289
Qwen3 8B (non-reasoning)
13.3
290
Gemini 2.5 Flash-Lite (non-reasoning)
13.3
291
Qwen3 8B
12.9
292
Gemma 3 27BFREE
12.9
293
Qwen3 Omni 30B A3B (指令)
12.3
294
DeepSeek R1 Distill Llama 70B
12.3
295
Ministral 3 8B
11.6
296
Claude 3 Haiku
11.3
297
Gemma 3 4BFREE
11.3
298
Gemma 3n E2BFREE
11.3
299
Gemma 4 E4B
11.2
300
Qwen3.5 0.8B
10.3
301
Qwen3.5 0.8B (non-reasoning)
10.3
302
Qwen3 1.7B
10.3
303
Qwen3 0.6B (non-reasoning)
10.3
304
Qwen3 1.7B (non-reasoning)
10.3
305
Qwen3 0.6B
10.3
306
Gemma 3 1BFREE
10.3
307
Gemma 3 12BFREE
6.5
308
Ministral 3 3B
6.1