Skip to main content
Radar

Every change to the models you run, with its source and its date. Releases, price changes, retirements, API changes, and incidents.Every change to the models you run, with its source.

Follow model changes
BenchLM

Artificial Analysis Omniscience Accuracy (AA-Omniscience Accuracy)

A display-only Artificial Analysis knowledge metric for the proportion of correctly answered questions.

Data verified 33 confirmed releases in the last 30 daysSee provider release alerts

Top models on AA-Omniscience Accuracy — September 15, 2026

As of September 15, 2026, Claude Fable 5.1 leads the AA-Omniscience Accuracy leaderboard with 67.2% , followed by Claude Fable 5 (65.4%) and GPT-6 Astra (62.6%).

1Proprietary

Claude Fable 5.1

Anthropic

67.2%
Overall 84.76Context 1M
2Proprietary

Claude Fable 5

Anthropic

65.4%
Overall 81.42Context 1M+
3Proprietary

GPT-6 Astra

OpenAI

62.6%
Overall 82.94Context 1.05M

190 modelsKnowledge6% of category scoreCurrentUpdated September 15, 2026

Leaderboard (190 models)

Score
1
Claude Fable 5.1Anthropic · Closed
67.2%
2
Claude Fable 5Anthropic · Closed
65.4%
3
GPT-6 AstraOpenAI · Closed
62.6%
4
Claude Opus 5Anthropic · Closed
60.9%
5
GPT-5.6 SolOpenAI · Closed
59.4%
6
GPT-5.5OpenAI · Closed
58.0%
7
Gemini 3 ProGoogle · Closed
55.8%
8
Gemini 3.7 FlashGoogle · Closed
55.3%
9
Gemini 3.1 ProGoogle · Closed
54.9%
10
Gemini 3.8 FlashGoogle · Closed
54.6%
11
GPT-5.3 CodexOpenAI · Closed
52.9%
12
GPT-5.3-Codex-SparkOpenAI · Closed
52.9%
13
Muse Spark 1.1Meta · Closed
52.1%
14
Gemini 3.5 FlashGoogle · Closed
51.9%
15
Grok 4.5xAI · Closed
51.6%
16
GPT-5.4OpenAI · Closed
50.8%
17
Gemini 3.6 FlashGoogle · Closed
50.0%
18
Muse SparkMeta · Closed
49.6%
19
DeepSeek V4 Pro 0813DeepSeek · Closed
49.1%
20
Claude Opus 4.7 (Adaptive)Anthropic · Closed
48.9%
21
Claude Opus 4.8Anthropic · Closed
48.8%
22
Grok 4.6xAI · Closed
48.2%
23
Kimi K3Moonshot AI · Closed
47.6%
24
Claude Opus 4.6 (Adaptive)Anthropic · Closed
47.0%
25
GPT-5.6 TerraOpenAI · Closed
46.8%
26
Claude Opus 4.5 ThinkingAnthropic · Closed
46.6%
27
DeepSeek V4.1 FlashDeepSeek · Open weight
46.4%
28
Claude Opus 4.6Anthropic · Closed
45.8%
29
Gemini 3 FlashGoogle · Closed
45.8%
30
Muse Spark 1.2Meta · Closed
45.4%
31
Claude Opus 4.7Anthropic · Closed
44.7%
32
GPT-5.2OpenAI · Closed
44.3%
33
Muse Spark 1.3Meta · Closed
43.6%
34
GPT-5.6 LunaOpenAI · Closed
42.7%
35
InklingThinking Machines Lab · Open weight
41.6%
36
DeepSeek V4 Pro (High)DeepSeek · Open weight
41.4%
37
GPT-5.2-CodexOpenAI · Closed
41.1%
38
Claude Opus 4.5Anthropic · Closed
40.9%
39
Grok 4xAI · Closed
40.5%
40
DeepSeek V4 Flash 0731DeepSeek · Closed
40.4%
41
GPT-5 (high)OpenAI · Closed
40.3%
42
Claude Sonnet 5Anthropic · Closed
40.1%
43
GPT-5.1-CodexOpenAI · Closed
39.9%
44
GPT-5.1-Codex-MaxOpenAI · Closed
39.9%
45
Kimi K2.7 CodeMoonshot AI · Open weight
39.6%
46
GPT-5 (medium)OpenAI · Closed
39.5%
47
Gemini 2.5 ProGoogle · Closed
39.1%
48
Claude Sonnet 4.6Anthropic · Closed
38.6%
49
o3OpenAI · Closed
38.6%
50
Qwen 3.6 Max (preview)Alibaba · Closed
37.9%
51
GPT-5.1OpenAI · Closed
37.7%
52
GPT-5.4 miniOpenAI · Closed
37.5%
53
Kimi K2.5Moonshot AI · Open weight
35.2%
54
Kimi K2.5 (Reasoning)Moonshot AI · Closed
35.2%
55
Grok 4.3xAI · Closed
34.6%
56
o1OpenAI · Closed
34.5%
57
GLM-5.3Z.AI · Open weight
33.9%
58
Inkling-SmallThinking Machines Lab · Open weight
33.2%
59
Kimi K2.6Moonshot AI · Open weight
32.6%
60
Hy3Tencent · Open weight
32.0%
61
Qwen3.8 Max PreviewAlibaba · Closed
31.9%
62
Apodex 1.1Apodex · Closed
31.7%
63
Apodex 1.1 MiniApodex · Open weight
31.7%
64
Hy3 PreviewTencent · Open weight
31.5%
65
Qwen3.7 MaxAlibaba · Closed
31.1%
66
DeepSeek-R1DeepSeek · Open weight
30.5%
67
Gemini 3.5 Flash-LiteGoogle · Closed
29.5%
68
GLM-4.7Z.AI · Open weight
29.3%
69
GLM-5V-TurboZ.AI · Closed
29.3%
70
DeepSeek V3.1 (Reasoning)DeepSeek · Open weight
29.0%
71
GLM-5-TurboZ.AI · Closed
28.4%
72
GPT-4.1OpenAI · Closed
27.8%
73
Kimi K2Moonshot AI · Closed
27.4%
74
Muse Glimmer 30BMeta · Open weight
27.0%
75
MiniMax M2.7MiniMax · Open weight
26.8%
76
MiMo-V2-ProXiaomi · Closed
26.6%
77
Qwen3.6 PlusAlibaba · Closed
26.4%
78
GLM-5Z.AI · Open weight
26.3%
79
MiniMax M2.5MiniMax · Closed
26.2%
80
Gemini 2.5 FlashGoogle · Closed
26.1%
81
Step 3.7 FlashStepFun · Open weight
25.8%
82
GPT-5.4 nanoOpenAI · Closed
25.7%
83
DeepSeek V3DeepSeek · Open weight
25.5%
84
Grok 4.1 Fast (Reasoning)xAI · Closed
25.1%
85
Mistral Large 3Mistral · Closed
25.0%
86
GPT-5 miniOpenAI · Closed
25.0%
87
Llama 4 MaverickMeta · Open weight
24.9%
88
Mistral Medium 3.5 128BMistral · Open weight
24.7%
89
Step 3.5 FlashStepFun · Open weight
24.6%
90
Qwen3.5 397BAlibaba · Open weight
24.5%
91
Qwen3.8-Flash-NextAlibaba · Open weight
24.5%
92
Qwen3.5 397B (Reasoning)Alibaba · Open weight
24.5%
93
Qwen3.5-122B-A10BAlibaba · Open weight
24.4%
94
Qwen3 MaxAlibaba · Closed
24.4%
95
GLM-5.2Z.AI · Open weight
24.3%
96
Nemotron 3 Super 100BNVIDIA · Open weight
24.3%
97
Nemotron 3 Super 120B A12BNVIDIA · Open weight
24.3%
98
DeepSeek V3.2DeepSeek · Open weight
24.0%
99
GLM-5.1Z.AI · Open weight
23.7%
100
Grok Code Fast 1xAI · Closed
23.5%
101
Llama 3.1 405BMeta · Open weight
23.2%
102
DeepSeek V3.1DeepSeek · Open weight
23.1%
103
Grok 4 Fast (Reasoning)xAI · Closed
22.8%
104
Claude 4 SonnetAnthropic · Closed
22.7%
105
Qwen3.7 PlusAlibaba · Closed
22.5%
106
Trinity-Large-ThinkingArcee AI · Open weight
22.5%
107
Trinity-Large-PreviewArcee AI · Open weight
22.5%
108
MiniMax M1 80kMiniMax · Closed
22.5%
109
MiMo-V2.5-ProXiaomi · Closed
22.4%
110
Mercury 2.5Inception · Closed
22.0%
111
GPT-OSS 120BOpenAI · Open weight
21.8%
112
Mistral Small 4Mistral · Open weight
21.7%
113
Mistral Small 4 (Reasoning)Mistral · Open weight
21.7%
114
Nemotron 3 UltraNVIDIA · Open weight
21.6%
115
GLM-4.6Z.AI · Open weight
21.4%
116
Mercury 2Inception · Closed
21.2%
117
Qwen3.5-27BAlibaba · Open weight
20.7%
118
GPT-4.1 miniOpenAI · Closed
20.3%
119
Qwen3.5-35B-A3BAlibaba · Open weight
20.1%
120
Nemotron Ultra 253BNVIDIA · Open weight
20.1%
121
Gemma 4 31BGoogle · Open weight
20.0%
122
GPT-4oOpenAI · Closed
19.9%
123
Mistral Large 2Mistral · Closed
19.9%
124
Qwen3.6-27BAlibaba · Open weight
19.6%
125
MiMo-V2-OmniXiaomi · Closed
19.3%
126
Gemma 4 26B A4BGoogle · Open weight
19.1%
127
GPT-5 nanoOpenAI · Closed
19.1%
128
Ultravox v0.6 Llama 3.3 70BFixie AI · Open weight
19.0%
129
Solar Pro 4Upstage · Closed
18.9%
130
North Mini CodeCohere · Open weight
18.9%
131
Qwen3.6-35B-A3BAlibaba · Open weight
18.8%
132
A.X K2SK Telecom · Open weight
18.6%
133
Solar Pro 3Upstage · Closed
18.5%
134
Mistral Medium 3Mistral · Closed
18.3%
135
Ling 3.0 FlashInclusionAI · Open weight
18.2%
136
Ling 3.0 Flash FP8InclusionAI · Open weight
18.2%
137
Claude 3 HaikuAnthropic · Closed
17.6%
138
Sarvam 105BSarvam · Open weight
17.6%
139
Nemotron 3 Nano 30BNVIDIA · Open weight
17.3%
140
Grok 4.1 FastxAI · Closed
17.2%
141
Nova ProAmazon · Closed
16.9%
142
MiniMax M3MiniMax · Open weight
16.7%
143
K-ExaoneLG AI Research · Closed
16.4%
144
GLM-4.5-AirZ.AI · Closed
16.3%
145
GLM-4.7-FlashZ.AI · Open weight
16.2%
146
Solar Pro 2Upstage · Closed
16.1%
147
GPT-OSS 20BOpenAI · Open weight
16.0%
148
Gemma 4 12BGoogle · Open weight
15.6%
149
Qwen3.8-27BAlibaba · Open weight
15.6%
150
MiMo-V2-FlashXiaomi · Open weight
15.6%
151
Ling 2.6 FlashInclusionAI · Open weight
15.6%
152
Quasar 438BMultiverse Computing · Closed
15.5%
153
Nemotron 3 Nano Omni 30B A3BNVIDIA · Open weight
15.2%
154
Llama 4 ScoutMeta · Open weight
15.2%
155
Qwen3-Omni-30B-A3B-ThinkingAlibaba · Open weight
14.6%
156
Ling 3.0 Flash VLInclusionAI · Open weight
14.4%
157
Nemotron 3.5 Lightning 30B A3B NVFP4NVIDIA · Open weight
14.4%
158
Qwen3-Omni-30B-A3B-InstructAlibaba · Open weight
14.3%
159
Phi-4Microsoft · Open weight
14.1%
160
GPT-4.1 nanoOpenAI · Closed
13.7%
161
Ministral 3 14B (Reasoning)Mistral · Open weight
13.6%
162
Ministral 3 14BMistral · Open weight
13.6%
163
K-EXAONE 2.0LG AI Research · Open weight
13.1%
164
Gemma 3 27BGoogle · Open weight
13.0%
165
Ministral 3 8B (Reasoning)Mistral · Open weight
13.0%
166
Ministral 3 8BMistral · Open weight
13.0%
167
Sarvam 30BSarvam · Open weight
12.6%
168
Granite 4.2 8BIBM · Open weight
11.2%
169
Celeris-1Celeris · Closed
11.0%
170
Exaone 4.0 32BLG AI Research · Open weight
10.6%
171
Granite 4.2 30BIBM · Open weight
10.1%
172
LFM2.5-8B-A1BLiquidAI · Open weight
9.4%
173
Granite 4.2 3BIBM · Open weight
9.2%
174
Ministral 3 3B (Reasoning)Mistral · Open weight
9.0%
175
Ministral 3 3BMistral · Open weight
9.0%
176
Command A+Cohere · Open weight
8.9%
177
Gemma 4 E4BGoogle · Open weight
8.6%
178
Ling 3.0 TinyInclusionAI · Open weight
8.5%
179
MiniCPM5-2BOpenBMB · Open weight
8.4%
180
LFM2.5-1.2B-ThinkingLiquidAI · Closed
7.4%
181
LFM2.5-1.2B-InstructLiquidAI · Closed
7.0%
182
Gemma 4 E2BGoogle · Open weight
6.6%
183
LFM2-24B-A2BLiquidAI · Closed
6.5%
184
Granite-4.0-1BIBM · Open weight
6.2%
185
LFM2.5-VL-1.6B-ExtractLiquidAI · Open weight
5.8%
186
Granite-4.0-H-1BIBM · Open weight
5.2%
187
Exaone 4.0 1.2BLG AI Research · Open weight
5.0%
188
LFM2.5-2.6BLiquidAI · Open weight
4.4%
189
Granite-4.0-350MIBM · Open weight
3.9%
190
Granite-4.0-H-350MIBM · Open weight
3.8%

According to BenchLM.ai, Claude Fable 5.1 leads the AA-Omniscience Accuracy benchmark with a score of 67.2%, followed by Claude Fable 5 (65.4%) and GPT-6 Astra (62.6%). The scores show moderate spread, with meaningful differences between the top tier and mid-tier models.

190 models have been evaluated on AA-Omniscience Accuracy. The benchmark falls in the Knowledge category. This category carries a 12% weight in BenchLM.ai's overall scoring system. Within that category, AA-Omniscience Accuracy contributes 6% of the category score, so strong performance here directly affects a model's overall ranking.

About AA-Omniscience Accuracy

Year

2026

Tasks

Knowledge questions

Format

Accuracy

Difficulty

Broad knowledge

BenchLM stores AA-Omniscience Accuracy as a display-only row when a model page publishes the exact Artificial Analysis benchmark card value.

Artificial Analysis model benchmarks

BenchLM freshness & provenance

Version

AA-Omniscience Accuracy 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

Current

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does AA-Omniscience Accuracy measure?

A display-only Artificial Analysis knowledge metric for the proportion of correctly answered questions.

Which model scores highest on AA-Omniscience Accuracy?

Claude Fable 5.1 by Anthropic currently leads with a score of 67.2% on AA-Omniscience Accuracy.

How many models are evaluated on AA-Omniscience Accuracy?

190 AI models have been evaluated on AA-Omniscience Accuracy on BenchLM.

Compare Top Models on AA-Omniscience Accuracy

Claude Fable 5.1 vs Claude Fable 5Claude Fable 5 vs GPT-6 AstraGPT-6 Astra vs Claude Opus 5Claude Opus 5 vs GPT-5.6 Sol
Last updated: September 15, 2026 · BenchLM version AA-Omniscience Accuracy 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.