Skip to main content
BenchLM
Advertise

Benchmark profile

Artificial Analysis Omniscience Accuracy (AA-Omniscience Accuracy)

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

Data verified

Benchmark score on AA-Omniscience Accuracy — July 23, 2026

BenchLM mirrors the published score view for AA-Omniscience Accuracy. Claude Fable 5 leads the public snapshot at 61.4% , followed by GPT-5.6 Sol (58.5%) and GPT-5.5 (56.9%). BenchLM does not use these results to rank models overall.

1Closed

Claude Fable 5

Anthropic

claude-fable-5

61.4%
Overall 83.68Context 1M+
2Closed

GPT-5.6 Sol

OpenAI

gpt-5-6-sol

58.5%
Overall 81.96Context 1M
3Closed

GPT-5.5

OpenAI

gpt-5-5

56.9%
Overall 73.51Context 1M
153 modelsKnowledgeCurrentDisplay onlyUpdated July 23, 2026

Benchmark score table (153 models)

Score
1
Claude Fable 5Anthropic · Closed
61.4%
2
GPT-5.6 SolOpenAI · Closed
58.5%
3
GPT-5.5OpenAI · Closed
56.9%
4
Gemini 3 ProGoogle · Closed
55.9%
5
Gemini 3.1 ProGoogle · Closed
55.3%
6
Grok 4.5xAI · Closed
52.1%
7
Gemini 3.5 FlashGoogle · Closed
51.9%
8
GPT-5.3 CodexOpenAI · Closed
51.8%
9
GPT-5.3-Codex-SparkOpenAI · Closed
51.8%
10
Gemini 3.6 FlashGoogle · Closed
50.2%
11
GPT-5.4OpenAI · Closed
50.0%
12
Claude Opus 4.8Anthropic · Closed
46.6%
13
Claude Opus 4.6 (Adaptive)Anthropic · Closed
46.4%
14
Kimi K3Moonshot AI · Closed
46.0%
15
GPT-5.6 TerraOpenAI · Closed
45.9%
16
Claude Opus 4.7 (Adaptive)Anthropic · Closed
45.8%
17
Claude Opus 4.5 ThinkingAnthropic · Closed
45.7%
18
Gemini 3 FlashGoogle · Closed
45.5%
19
Claude Opus 4.6Anthropic · Closed
45.2%
20
Muse SparkMeta · Closed
44.6%
21
GPT-5.2OpenAI · Closed
43.8%
22
Claude Opus 4.7Anthropic · Closed
43.5%
23
DeepSeek V4 Pro (Max)DeepSeek · Open weight
43.3%
24
DeepSeek V4 Pro (High)DeepSeek · Open weight
41.8%
25
GPT-5.6 LunaOpenAI · Closed
41.5%
26
Grok 4xAI · Closed
41.4%
27
Claude Opus 4.5Anthropic · Closed
40.7%
28
GPT-5 (high)OpenAI · Closed
40.7%
29
GPT-5.2-CodexOpenAI · Closed
40.7%
30
Muse Spark 1.1Meta · Closed
40.6%
31
InklingThinking Machines Lab · Open weight
40.0%
32
GPT-5.1-Codex-MaxOpenAI · Closed
39.2%
33
GPT-5.1-CodexOpenAI · Closed
39.2%
34
Gemini 2.5 ProGoogle · Closed
39.0%
35
GPT-5 (medium)OpenAI · Closed
38.9%
36
Kimi K2.7 CodeMoonshot AI · Open weight
38.6%
37
o3OpenAI · Closed
38.4%
38
Claude Sonnet 5Anthropic · Closed
38.3%
39
Claude Sonnet 4.6Anthropic · Closed
38.0%
40
Qwen 3.6 Max (preview)Alibaba · Closed
37.7%
41
GPT-5.1OpenAI · Closed
37.6%
42
GPT-5.4 miniOpenAI · Closed
37.5%
43
DeepSeek V4 Flash (Max)DeepSeek · Open weight
37.2%
44
Gemini 3.1 Flash-LiteGoogle · Closed
36.4%
45
DeepSeek V4 Flash (High)DeepSeek · Open weight
35.5%
46
o1OpenAI · Closed
34.7%
47
Grok 4.3xAI · Closed
34.6%
48
Kimi K2.5Moonshot AI · Open weight
34.3%
49
Kimi K2.5 (Reasoning)Moonshot AI · Closed
34.3%
50
Kimi K2.6Moonshot AI · Open weight
32.8%
51
Hy3 PreviewTencent · Open weight
31.5%
52
Hy3Tencent · Open weight
31.5%
53
Qwen3.5 397BAlibaba · Open weight
31.4%
54
Qwen3.5 397B (Reasoning)Alibaba · Open weight
31.4%
55
DeepSeek-R1DeepSeek · Open weight
31.0%
56
Gemini 3.5 Flash-LiteGoogle · Closed
30.3%
57
Qwen3.7 MaxAlibaba · Closed
30.1%
58
GLM-4.7Z.AI · Open weight
29.3%
59
GLM-5V-TurboZ.AI · Closed
29.1%
60
GLM-5-TurboZ.AI · Closed
29.0%
61
DeepSeek V3.1 (Reasoning)DeepSeek · Open weight
28.8%
62
GLM-5Z.AI · Open weight
26.9%
63
Kimi K2Moonshot AI · Closed
26.8%
64
MiMo-V2-ProXiaomi · Closed
26.8%
65
Gemini 2.5 FlashGoogle · Closed
26.5%
66
Qwen3.6 PlusAlibaba · Closed
26.2%
67
MiniMax M2.5MiniMax · Closed
26.2%
68
MiniMax M2.7MiniMax · Open weight
26.1%
69
GPT-5.4 nanoOpenAI · Closed
25.4%
70
Step 3.7 FlashStepFun · Open weight
25.4%
71
DeepSeek V3DeepSeek · Open weight
25.4%
72
Grok 4.1 Fast (Reasoning)xAI · Closed
25.3%
73
GLM-5.2Z.AI · Open weight
25.1%
74
Mistral Medium 3.5 128BMistral · Open weight
25.1%
75
Step 3.5 FlashStepFun · Open weight
25.0%
76
Qwen3.5-122B-A10BAlibaba · Open weight
24.7%
77
Qwen3 MaxAlibaba · Closed
24.4%
78
Llama 4 MaverickMeta · Open weight
24.3%
79
GLM-5.1Z.AI · Open weight
24.2%
80
DeepSeek V3.2DeepSeek · Open weight
24.2%
81
GPT-4.1OpenAI · Closed
24.2%
82
Mistral Large 3Mistral · Closed
24.1%
83
GPT-5 miniOpenAI · Closed
24.0%
84
Nemotron 3 Super 120B A12BNVIDIA · Open weight
24.0%
85
Grok Code Fast 1xAI · Closed
23.8%
86
DeepSeek V3.1DeepSeek · Open weight
23.1%
87
Trinity-Large-PreviewArcee AI · Open weight
22.8%
88
Trinity-Large-ThinkingArcee AI · Open weight
22.8%
89
MiMo-V2.5-ProXiaomi · Closed
22.6%
90
Grok 4 Fast (Reasoning)xAI · Closed
22.6%
91
Claude 4 SonnetAnthropic · Closed
22.4%
92
Llama 3.1 405BMeta · Open weight
22.3%
93
Qwen3.7 PlusAlibaba · Closed
22.2%
94
Mistral Small 4Mistral · Open weight
22.1%
95
Mistral Small 4 (Reasoning)Mistral · Open weight
22.1%
96
Nemotron 3 UltraNVIDIA · Open weight
21.6%
97
GPT-OSS 120BOpenAI · Open weight
21.5%
98
MiniMax M1 80kMiniMax · Closed
21.1%
99
Qwen3.5-27BAlibaba · Open weight
21.0%
100
GLM-4.6Z.AI · Open weight
20.8%
101
Qwen3.5-35B-A3BAlibaba · Open weight
20.5%
102
Mercury 2Inception · Closed
20.5%
103
Mistral Large 2Mistral · Closed
20.1%
104
Gemma 4 31BGoogle · Open weight
19.9%
105
Nemotron Ultra 253BNVIDIA · Open weight
19.9%
106
GPT-4oOpenAI · Closed
19.7%
107
Qwen3.6-27BAlibaba · Open weight
19.2%
108
Qwen3.6-35B-A3BAlibaba · Open weight
18.9%
109
MiMo-V2-OmniXiaomi · Closed
18.7%
110
Mistral Medium 3Mistral · Closed
18.3%
111
GPT-5 nanoOpenAI · Closed
18.3%
112
Gemma 4 26B A4BGoogle · Open weight
18.2%
113
Sarvam 105BSarvam · Open weight
17.6%
114
GPT-4.1 miniOpenAI · Closed
17.5%
115
Claude 3 HaikuAnthropic · Closed
17.2%
116
Nemotron 3 Nano 30BNVIDIA · Open weight
17.1%
117
Grok 4.1 FastxAI · Closed
17.0%
118
Nova ProAmazon · Closed
17.0%
119
K-ExaoneLG AI Research · Closed
16.5%
120
Gemma 4 12BGoogle · Open weight
16.0%
121
GLM-4.7-FlashZ.AI · Open weight
15.9%
122
Solar Pro 2Upstage · Closed
15.6%
123
GLM-4.5-AirZ.AI · Closed
15.5%
124
GPT-OSS 20BOpenAI · Open weight
15.5%
125
Ling 2.6 FlashInclusionAI · Open weight
15.4%
126
MiMo-V2-FlashXiaomi · Open weight
15.2%
127
MiniMax M3MiniMax · Open weight
15.0%
128
Nemotron 3 Nano Omni 30B A3BNVIDIA · Open weight
14.8%
129
Llama 4 ScoutMeta · Open weight
14.6%
130
GPT-4.1 nanoOpenAI · Closed
13.3%
131
Phi-4Microsoft · Open weight
13.2%
132
Sarvam 30BSarvam · Open weight
12.7%
133
Gemma 3 27BGoogle · Open weight
12.5%
134
Ministral 3 14B (Reasoning)Mistral · Open weight
12.3%
135
Ministral 3 14BMistral · Open weight
12.3%
136
Ministral 3 8B (Reasoning)Mistral · Open weight
11.2%
137
Ministral 3 8BMistral · Open weight
11.2%
138
Exaone 4.0 32BLG AI Research · Open weight
10.4%
139
LFM2.5-8B-A1BLiquidAI · Open weight
9.4%
140
Command A+Cohere · Open weight
8.9%
141
Gemma 4 E4BGoogle · Open weight
8.6%
142
Ministral 3 3B (Reasoning)Mistral · Open weight
7.6%
143
Ministral 3 3BMistral · Open weight
7.6%
144
Gemma 4 E2BGoogle · Open weight
6.7%
145
LFM2.5-1.2B-ThinkingLiquidAI · Closed
6.6%
146
LFM2-24B-A2BLiquidAI · Closed
6.4%
147
Granite-4.0-1BIBM · Open weight
6.1%
148
LFM2.5-1.2B-InstructLiquidAI · Closed
6.0%
149
Granite-4.0-H-1BIBM · Open weight
5.3%
150
LFM2.5-VL-1.6B-ExtractLiquidAI · Open weight
5.2%
151
Exaone 4.0 1.2BLG AI Research · Open weight
4.7%
152
Granite-4.0-H-350MIBM · Open weight
3.7%
153
Granite-4.0-350MIBM · Open weight
3.2%

The published AA-Omniscience Accuracy snapshot places Claude Fable 5 first at 61.4%. The third row is 4.5 points behind. The broader top-10 range is 11.2 points, so the table still separates the published systems.

153 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. AA-Omniscience Accuracy is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

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

CurrentDisplay only

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 by Anthropic currently leads with a score of 61.4% on AA-Omniscience Accuracy.

How many models are evaluated on AA-Omniscience Accuracy?

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

Compare Top Models on AA-Omniscience Accuracy

Claude Fable 5 vs GPT-5.6 SolGPT-5.6 Sol vs GPT-5.5GPT-5.5 vs Gemini 3 ProGemini 3 Pro vs Gemini 3.1 Pro
Last updated: July 23, 2026 · BenchLM version AA-Omniscience Accuracy 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.