Clarity for AI buildersSponsored by AI Routing
HomeRankings
Model Index · AI Routing & independent evals

AI Model Leaderboard

Scores combine AI Routing measurements with results from independent third-party evaluators, normalized to a single scale and weighted by domain.

698
Snapshot dates2026-10-26T06:55:22.587Z—

Model rankings

Choose a category. Each view uses its own source dataset; periods are available for Top Models and Leaderboard.

698
#ModelSite ScoreData coverage
01
Claude Fable 5.1 (Max, Default Fallback)Anthropic
67.580% · 2/3
02
Claude Fable 5.1 (Xhigh, Default Fallback)Anthropic
67.080% · 2/3
03
Claude Fable 5.1 (High, Default Fallback)Anthropic
65.280% · 2/3
04
GPT-6 Astra (Max)OpenAI
64.880% · 2/3
05
Claude Opus 5 (Max)Anthropic
64.480% · 2/3
06
GPT-6 Astra (Xhigh)OpenAI
64.280% · 2/3
07
GPT-6 Astra (High)OpenAI
64.080% · 2/3
08
Claude Opus 5 (Xhigh)Anthropic
63.480% · 2/3
09
GPT-6 Astra (Medium)OpenAI
63.280% · 2/3
10
Claude Fable 5 (Max, Opus 4.8 Fallback)Anthropic
63.080% · 2/3
11
Claude Fable 5.1 (Medium, Default Fallback)Anthropic
63.080% · 2/3
12
Claude Opus 5 (High)Anthropic
62.380% · 2/3
13
GPT-5.6 Sol (Max)OpenAI
62.280% · 2/3
14
Muse Spark 1.3 (Max)Meta
62.080% · 2/3
15
GPT-5.6 Sol (Xhigh)OpenAI
61.180% · 2/3
16
Claude Fable 5.1 (Low, Default Fallback)Anthropic
61.080% · 2/3
17
Muse Spark 1.3 (Xhigh)Meta
60.880% · 2/3
18
Qwen3.8 Max (0902)Alibaba
60.880% · 2/3
19
GPT-6 Astra (Low)OpenAI
60.880% · 2/3
20
Grok 4.6 (High)SpaceXAI
60.580% · 2/3
21
Grok 4.6 (Xhigh)SpaceXAI
60.180% · 2/3
22
Kimi K3 (Max)Kimi
59.980% · 2/3
23
GLM-5.3 (Max)Z AI
59.880% · 2/3
24
GPT-5.6 Sol (High)OpenAI
59.880% · 2/3
25
Claude Opus 5 (Medium)Anthropic
59.580% · 2/3
26
GPT-5.6 Terra (Max)OpenAI
59.480% · 2/3
27
Grok 4.6 (Medium)SpaceXAI
58.680% · 2/3
28
Gemini 3.8 Flash (High)Google
58.680% · 2/3
29
Claude Opus 4.8 (Max)Anthropic
58.080% · 2/3
30
GPT-5.6 Sol (Medium)OpenAI
57.880% · 2/3
31
Claude Opus 5.5 (Max, Default Fallback)Anthropic
57.640% · 1/3
32
Gemini 3.7 Flash (High)Google
57.680% · 2/3
33
Claude Opus 4.7 (Max)Anthropic
57.180% · 2/3
34
Gemini 3.8 Flash (Medium)Google
56.980% · 2/3
35
GPT-5.5 (Xhigh)OpenAI
56.780% · 2/3
36
GLM-5.3-FlashZ AI
56.680% · 2/3
37
Qwen3.8-Flash-NextAlibaba
56.480% · 2/3
38
Qwen3.8 MaxAlibaba
56.080% · 2/3
39
Claude Opus 5.5 (Xhigh, Default Fallback)Anthropic
56.040% · 1/3
40
Claude Sonnet 5.5 (Max, Default Fallback)Anthropic
56.040% · 1/3
41
Muse Spark 1.2 (Xhigh)Meta
55.980% · 2/3
42
Qwen3.8 2.4T A95BAlibaba
55.980% · 2/3
43
Muse Wonder Full (Max)Meta
55.840% · 1/3
44
Grok 4.5 (High)SpaceXAI
55.680% · 2/3
45
Gemini 3.7 Flash (Medium)Google
55.580% · 2/3
46
Kimi K3.1 ProKimi
55.140% · 1/3
47
GPT-5.4 (Xhigh)OpenAI
55.080% · 2/3
48
Claude Sonnet 5 (Max)Anthropic
54.980% · 2/3
49
GPT-5.6 Luna (Max)OpenAI
54.480% · 2/3
50
GPT-5.5 (High)OpenAI
54.380% · 2/3
51
GPT-5.6 Terra (Xhigh)OpenAI
54.380% · 2/3
52
Gemini 3.7 Flash (Low)Google
54.080% · 2/3
53
Claude Opus 5.5 (High, Default Fallback)Anthropic
53.640% · 1/3
54
Gemini 3.8 Flash (Low)Google
53.580% · 2/3
55
Claude Opus 5 (Low)Anthropic
53.280% · 2/3
56
GPT-5.5 (Medium)OpenAI
52.680% · 2/3
57
Gemini 4 Argon (High)Google
52.640% · 1/3
58
Muse Spark 1.1 (Xhigh)Meta
52.580% · 2/3
59
DeepSeek V4 Pro 0813 (Max)DeepSeek
52.480% · 2/3
60
Muse Wonder (Max)Meta
52.440% · 1/3
61
Claude Sonnet 5.5 (Xhigh, Default Fallback)Anthropic
51.940% · 1/3
62
GPT-6.1 Sol (Max)OpenAI
51.840% · 1/3
63
DeepSeek V4 Flash 0731 (Max)DeepSeek
51.780% · 2/3
64
Gemini 3.6 Flash (High)Google
51.680% · 2/3
65
GPT-5.6 Sol (Low)OpenAI
51.680% · 2/3
66
GPT-5.6 Luna (Xhigh)OpenAI
51.680% · 2/3
67
Gemini 3.5 Flash (High)Google
51.380% · 2/3
68
GLM-5.2 (Max)Z AI
51.380% · 2/3
69
Claude Opus 5.5 (Medium, Default Fallback)Anthropic
51.240% · 1/3
70
Kimi K3 (Low)Kimi
51.080% · 2/3
71
GPT-6.1 Sol (Xhigh)OpenAI
51.040% · 1/3
72
Qwen3.8 27B (Xhigh)Alibaba
50.980% · 2/3
73
Grok 4.6 (Low)SpaceXAI
50.780% · 2/3
74
GPT-5.6 Terra (High)OpenAI
50.680% · 2/3
75
GPT-6.1 Sol (High)OpenAI
50.240% · 1/3
76
DeepSeek V4 Flash Vision (Max)DeepSeek
49.980% · 2/3
77
Gemini 3.1 Pro PreviewGoogle
49.380% · 2/3
78
Agnes 2.5 Pro BetaSapiens AI
48.880% · 2/3
79
Motif 3Motif Technologies
48.580% · 2/3
80
GPT-6.1 Sol (Medium)OpenAI
47.840% · 1/3
81
Qwen3.7 MaxAlibaba
47.880% · 2/3
82
GPT-5.6 Luna (High)OpenAI
47.780% · 2/3
83
GPT-6 Sol (Max)OpenAI
47.640% · 1/3
84
GPT-5.6 Terra (Medium)OpenAI
47.480% · 2/3
85
Motif 3 (Beta)Motif Technologies
47.180% · 2/3
86
Claude Sonnet 5.5 (High, Default Fallback)Anthropic
46.840% · 1/3
87
GPT-5.6 Sol (Non-reasoning)OpenAI
46.780% · 2/3
88
Claude Sonnet 4.6 (Max)Anthropic
46.580% · 2/3
89
Grok 4.7 (Xhigh)SpaceXAI
46.440% · 1/3
90
Grok 4.7 (High)SpaceXAI
46.340% · 1/3
91
MiMo-V2.6-ProXiaomi
46.340% · 1/3
92
K2 Horizon 375B A23BInstitute of Foundation Models
46.080% · 2/3
93
GPT-5.5 (Low)OpenAI
45.880% · 2/3
94
DeepSeek V4.1 Flash 1017 (Max)DeepSeek
45.140% · 1/3
95
Muse SparkMeta
45.080% · 2/3
96
DeepSeek V4 Pro 0424 (Max)DeepSeek
44.980% · 2/3
97
Claude Sonnet 5 (Non-reasoning, High)Anthropic
44.880% · 2/3
98
DeepSeek V4 Pro 0424 (High)DeepSeek
44.480% · 2/3
99
Kimi K2.6 (Reasoning)Kimi
44.480% · 2/3
100
GPT-6 Sol (Xhigh)OpenAI
44.240% · 1/3
101
Quasar 438B (Max, Based on GLM-5.2)Multiverse Computing
44.080% · 2/3
102
MiniMax-M3MiniMax
43.980% · 2/3
103
Ling 3.1 Flash 1019InclusionAI
43.940% · 1/3
104
Step 5 PreviewStepFun
43.740% · 1/3
105
Nex-N2-Pro (Based on Qwen3.5-397B-A17B)Nex AGI
43.680% · 2/3
106
Apodex 1.1Apodex
43.680% · 2/3
107
Claude Haiku 5.5 (Max)Anthropic
43.440% · 1/3
108
Kimi K2.7 CodeKimi
43.380% · 2/3
109
MiMo-V2.5-Pro (Reasoning)Xiaomi
43.180% · 2/3
110
GPT-5.6 Terra (Low)OpenAI
42.880% · 2/3
111
MiniMax-M3.1-FlashMiniMax
42.740% · 1/3
112
GPT-6 Sol (High)OpenAI
42.440% · 1/3
113
Claude Opus 5.5 (Low, Default Fallback)Anthropic
42.340% · 1/3
114
Qwen3.8 27B (Low)Alibaba
42.280% · 2/3
115
Grok 4.7 (Low)SpaceXAI
42.240% · 1/3
116
GPT-6.1 Sol (Low)OpenAI
42.140% · 1/3
117
Hy3Tencent
42.080% · 2/3
118
Qwen3.8 27B (Medium)Alibaba
41.980% · 2/3
119
Claude Haiku 5.5 (Xhigh)Anthropic
41.240% · 1/3
120
Ling 3.1 FlashInclusionAI
41.140% · 1/3
121
GLM-5.1 (Reasoning)Z AI
41.080% · 2/3
122
Ling-3.0-flash-VLInclusionAI
40.880% · 2/3
123
Claude Sonnet 5.5 (Medium, Default Fallback)Anthropic
40.840% · 1/3
124
Qwen3.6 PlusAlibaba
40.880% · 2/3
125
KAT Coder Pro V2KwaiKAT
40.680% · 2/3
126
Qwen3.7 PlusAlibaba
40.580% · 2/3
127
DeepSeek V4 Flash 0420 (Max)DeepSeek
40.280% · 2/3
128
GPT-5.4 mini (Xhigh)OpenAI
40.180% · 2/3
129
GPT-5.5 (Non-reasoning)OpenAI
39.980% · 2/3
130
JT-4.1 Flash 236B A21B (Non-reasoning)China Mobile
39.980% · 2/3
131
GPT-6 Sol (Medium)OpenAI
39.840% · 1/3
132
DeepSeek V4.1 Flash (Max)DeepSeek
39.540% · 1/3
133
Grok Build 0.1 0616SpaceXAI
39.480% · 2/3
134
Inkling SmallThinking Machines
39.380% · 2/3
135
MiMo-V2.5Xiaomi
39.380% · 2/3
136
Ling-3.0-flash-FinInclusionAI
39.180% · 2/3
137
Muse Spark Open (Xhigh)Meta
38.940% · 1/3
138
Inkling (Xhigh)Thinking Machines
38.580% · 2/3
139
GPT-5.4 nano (Xhigh)OpenAI
38.480% · 2/3
140
Mistral Large 4 PreviewMistral
38.440% · 1/3
141
DeepSeek V4 Flash 0420 (High)DeepSeek
38.280% · 2/3
142
GPT-6 Luna (Max)OpenAI
38.140% · 1/3
143
MiMo-V2.6-FlashXiaomi
37.940% · 1/3
144
GPT-5.6 Luna (Medium)OpenAI
37.980% · 2/3
145
Claude Haiku 5.5 (High)Anthropic
37.840% · 1/3
146
MiniMax-M2.7MiniMax
37.780% · 2/3
147
Qwen3.6 27B (Reasoning)Alibaba
37.580% · 2/3
148
GPT-5.1 (High)OpenAI
37.080% · 2/3
149
GPT-5.6 Terra (Non-reasoning)OpenAI
36.580% · 2/3
150
Claude 4.5 Sonnet (Reasoning)Anthropic
36.480% · 2/3
151
Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA
36.180% · 2/3
152
Claude Sonnet 5.5 (Low, Default Fallback)Anthropic
35.940% · 1/3
153
Gemini 3.5 Flash-LiteGoogle
35.880% · 2/3
154
Ling 3.0 FlashInclusionAI
35.480% · 2/3
155
Kimi K2.5 (Reasoning)Kimi
35.180% · 2/3
156
Solar Open2 250BUpstage
34.980% · 2/3
157
GPT-6 Luna (Xhigh)OpenAI
34.640% · 1/3
158
Claude Haiku 5.5 (Medium)Anthropic
34.540% · 1/3
159
GLM-5.2 (Non-reasoning)Z AI
34.580% · 2/3
160
Claude Sonnet 5 (Xhigh)Anthropic
34.440% · 1/3
161
GLM-5.3 (Low)Z AI
34.340% · 1/3
162
GPT-6 Sol (Low)OpenAI
34.240% · 1/3
163
JT-4.1 Flash 236B A21B (Reasoning)China Mobile
33.940% · 1/3
164
GLM-4.7 (Reasoning)Z AI
33.880% · 2/3
165
Gemini 3.5 Flash (Medium)Google
33.640% · 1/3
166
Grok 4.3 (High)SpaceXAI
33.580% · 2/3
167
Qwen3.5 397B A17B (Reasoning)Alibaba
33.380% · 2/3
168
Muse Glimmer (High)Meta
33.380% · 2/3
169
Qwen3.6 27B (Non-reasoning)Alibaba
33.280% · 2/3
170
MiMo-V2-Flash (Non-reasoning)Xiaomi
32.980% · 2/3
171
GPT-6 Luna (High)OpenAI
32.940% · 1/3
172
DeepSeek V3.2 (Reasoning)DeepSeek
32.980% · 2/3
173
GPT-5.5 Instant (June 2026)OpenAI
32.780% · 2/3
174
GPT-5.6 Luna (Low)OpenAI
32.680% · 2/3
175
GPT-5.3 Codex (Xhigh)OpenAI
32.540% · 1/3
176
Qwen3.8 27B (Non-reasoning)Alibaba
32.480% · 2/3
177
LongCat 2.0LongCat
32.280% · 2/3
178
GLM-4.6 (Reasoning)Z AI
32.180% · 2/3
179
Claude Opus 4.6 (Max)Anthropic
31.940% · 1/3
180
Claude Sonnet 5 (High)Anthropic
31.740% · 1/3
181
Claude Opus 4.7 (Non-reasoning, High)Anthropic
30.940% · 1/3
182
Gemini 2.5 Pro Preview (Mar' 25)Google
30.980% · 2/3
183
Qwen3.5 122B A10B (Reasoning)Alibaba
30.780% · 2/3
184
Mistral Medium 3.5Mistral
30.580% · 2/3
185
Qwen3.5 122B A10B (Non-reasoning)Alibaba
30.580% · 2/3
186
Claude 4.5 Haiku (Reasoning)Anthropic
30.480% · 2/3
187
GPT-5 (High)OpenAI
30.480% · 2/3
188
GPT-5.2 (Xhigh)OpenAI
30.440% · 1/3
189
Beam 1.0Reflection
30.340% · 1/3
190
A.X-K2SK Telecom
30.180% · 2/3
191
K-EXAONE 2.0 0803LG AI Research
30.180% · 2/3
192
Qwen3.6 35B A3B (Reasoning)Alibaba
30.080% · 2/3
193
GPT-6 Luna (Medium)OpenAI
29.940% · 1/3
194
Ring-2.6-1TInclusionAI
29.780% · 2/3
195
K2 Horizon 7BInstitute of Foundation Models
29.680% · 2/3
196
Step 3.7 FlashStepFun
29.680% · 2/3
197
Claude Haiku 5.5 (Low)Anthropic
29.540% · 1/3
198
DeepSeek V3.1 Terminus (Reasoning)DeepSeek
29.180% · 2/3
199
Claude Opus 4.5 (Reasoning)Anthropic
29.140% · 1/3
200
Gemma 4 31B (Reasoning)Google
29.080% · 2/3

Source: Routing Brief · AA dataRouting Brief weighted score · rb-weighted-v1. Intelligence 40%, Coding 40%, Agentic 20%; missing weights redistributed. Coverage is not confidence. Source: Artificial Analysis.Sources & methodology

Missing data is shown as unavailable, never as zero. Report a data correction.