AA-Omniscience Accuracy leaderboard

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

Models ranked by their published Artificial Analysis Omniscience Accuracy value

Rank	Model	Provider	Accuracy
1	Claude Fable 5.1	Anthropic	67.2
2	Claude Fable 5	Anthropic	65.4
3	GPT-6 Astra	OpenAI	62.6
4	Claude Opus 5	Anthropic	60.9
5	GPT-5.6 Sol	OpenAI	59.4
6	GPT-5.5	OpenAI	58.0
7	Gemini 3 Pro	Google	55.8
8	Gemini 3.7 Flash	Google	55.3
9	Gemini 3.1 Pro	Google	54.9
10	Gemini 3.8 Flash	Google	54.6
11	GPT-5.3 Codex	OpenAI	52.9
12	Muse Spark 1.1	Meta	52.1
13	Gemini 3.5 Flash	Google	51.9
14	Grok 4.5	xAI	51.6
15	GPT-5.4	OpenAI	50.8
16	Gemini 3.6 Flash	Google	50.0
17	Muse Spark	Meta	49.6
18	DeepSeek V4 Pro 0813	DeepSeek	49.1
19	Claude Opus 4.7 (Adaptive)	Anthropic	48.9
20	Claude Opus 4.8	Anthropic	48.8
21	Grok 4.6	xAI	48.2
22	Kimi K3	Moonshot AI	47.6
23	Claude Opus 4.6 (Adaptive)	Anthropic	47.0
24	GPT-5.6 Terra	OpenAI	46.8
25	Claude Opus 4.5 Thinking	Anthropic	46.6
26	DeepSeek V4.1 Flash	DeepSeek	46.4
27	Claude Opus 4.6	Anthropic	45.8
27	Gemini 3 Flash	Google	45.8
29	Muse Spark 1.2	Meta	45.4
30	Claude Opus 4.7	Anthropic	44.7
31	GPT-5.2	OpenAI	44.3
32	Muse Spark 1.3	Meta	43.6
33	GPT-5.6 Luna	OpenAI	42.7
34	TMInkling	Thinking Machines Lab	41.6
35	GPT-5.2-Codex	OpenAI	41.1
36	Claude Opus 4.5	Anthropic	40.9
37	Grok 4	xAI	40.5
38	DeepSeek V4 Flash 0731	DeepSeek	40.4
39	GPT-5 (high)	OpenAI	40.3
40	Claude Sonnet 5	Anthropic	40.1
41	GPT-5.1-Codex	OpenAI	39.9
41	GPT-5.1-Codex-Max	OpenAI	39.9
43	Kimi K2.7 Code	Moonshot AI	39.6
44	GPT-5 (medium)	OpenAI	39.5
45	Gemini 2.5 Pro	Google	39.1
46	Claude Sonnet 4.6	Anthropic	38.6
46	o3	OpenAI	38.6
48	Qwen 3.6 Max (preview)	Alibaba	37.9
49	GPT-5.1	OpenAI	37.7
50	GPT-5.4 mini	OpenAI	37.5
51	Kimi K2.5	Moonshot AI	35.2
51	Kimi K2.5 (Reasoning)	Moonshot AI	35.2
53	Grok 4.3	xAI	34.6
54	o1	OpenAI	34.5
55	GLM-5.3	Z.AI	33.9
56	TMInkling-Small	Thinking Machines Lab	33.2
57	Kimi K2.6	Moonshot AI	32.6
58	Hy3	Tencent	32.0
59	APApodex 1.1	Apodex	31.7
59	APApodex 1.1 Mini	Apodex	31.7
59	Qwen3.8 Max Preview	Alibaba	31.7
62	Hy3 Preview	Tencent	31.5
63	Qwen3.7 Max	Alibaba	31.1
64	DeepSeek-R1	DeepSeek	30.5
65	Gemini 3.5 Flash-Lite	Google	29.5
66	GLM-4.7	Z.AI	29.3
66	GLM-5V-Turbo	Z.AI	29.3
68	DeepSeek V3.1 (Reasoning)	DeepSeek	29.0
69	GLM-5-Turbo	Z.AI	28.4
70	GPT-4.1	OpenAI	27.8
71	Kimi K2	Moonshot AI	27.4
72	Muse Glimmer 30B	Meta	27.0
73	MiniMax M2.7	MiniMax	26.8
74	MiMo-V2-Pro	Xiaomi	26.6
75	Qwen3.6 Plus	Alibaba	26.4
76	GLM-5	Z.AI	26.3
77	Gemini 2.5 Flash	Google	26.1
78	Step 3.7 Flash	StepFun	25.8
79	GPT-5.4 nano	OpenAI	25.7
80	DeepSeek V3	DeepSeek	25.5
81	Grok 4.1 Fast (Reasoning)	xAI	25.1
82	Mistral Large 3	Mistral	25.0
83	Llama 4 Maverick	Meta	24.9
84	Mistral Medium 3.5 128B	Mistral	24.7
85	Qwen3.5 397B	Alibaba	24.5
85	Qwen3.5 397B (Reasoning)	Alibaba	24.5
85	Qwen3.8-Flash-Next	Alibaba	24.5
88	Qwen3 Max	Alibaba	24.4
88	Qwen3.5-122B-A10B	Alibaba	24.4
90	GLM-5.2	Z.AI	24.3
90	Nemotron 3 Super 100B	NVIDIA	24.3
92	DeepSeek V3.2	DeepSeek	24.0
93	GLM-5.1	Z.AI	23.7
94	Grok Code Fast 1	xAI	23.5
95	Llama 3.1 405B	Meta	23.2
96	DeepSeek V3.1	DeepSeek	23.1
97	Grok 4 Fast (Reasoning)	xAI	22.8
98	Claude 4 Sonnet	Anthropic	22.7
99	Qwen3.7 Plus	Alibaba	22.5
99	Trinity-Large-Preview	Arcee AI	22.5
99	Trinity-Large-Thinking	Arcee AI	22.5
102	MiMo-V2.5-Pro	Xiaomi	22.4
103	Mercury 2.5	Inception	22.0
104	GPT-OSS 120B	OpenAI	21.8
105	Mistral Small 4	Mistral	21.7
105	Mistral Small 4 (Reasoning)	Mistral	21.7
107	Nemotron 3 Ultra	NVIDIA	21.6
108	GLM-4.6	Z.AI	21.4
109	Qwen3.5-27B	Alibaba	20.7
110	GPT-4.1 mini	OpenAI	20.3
111	Nemotron Ultra 253B	NVIDIA	20.1
111	Qwen3.5-35B-A3B	Alibaba	20.1
113	Gemma 4 31B	Google	20.0
114	GPT-4o	OpenAI	19.9
114	Mistral Large 2	Mistral	19.9
116	Qwen3.6-27B	Alibaba	19.6
117	MiMo-V2-Omni	Xiaomi	19.3
118	Gemma 4 26B A4B	Google	19.1
119	FAUltravox v0.6 Llama 3.3 70B	Fixie AI	19.0
120	North Mini Code	Cohere	18.9
120	Solar Pro 4	Upstage	18.9
122	Qwen3.6-35B-A3B	Alibaba	18.8
123	STA.X K2	SK Telecom	18.6
124	Solar Pro 3	Upstage	18.5
125	Mistral Medium 3	Mistral	18.3
126	Ling 3.0 Flash	InclusionAI	18.2
126	Ling 3.0 Flash FP8	InclusionAI	18.2
128	Claude 3 Haiku	Anthropic	17.6
128	SASarvam 105B	Sarvam	17.6
130	Nemotron 3 Nano 30B	NVIDIA	17.3
131	Grok 4.1 Fast	xAI	17.2
132	Nova Pro	Amazon	16.9
133	MiniMax M3	MiniMax	16.7
134	K-Exaone	LG AI Research	16.4
135	GLM-4.5-Air	Z.AI	16.3
136	Solar Pro 2	Upstage	16.1
137	GPT-OSS 20B	OpenAI	16.0
138	Gemma 4 12B	Google	15.6
138	Ling 2.6 Flash	InclusionAI	15.6
138	MiMo-V2-Flash	Xiaomi	15.6
138	Qwen3.8-27B	Alibaba	15.6
142	MCQuasar 438B	Multiverse Computing	15.5
143	Llama 4 Scout	Meta	15.2
143	Nemotron 3 Nano Omni 30B A3B	NVIDIA	15.2
145	Qwen3-Omni-30B-A3B-Thinking	Alibaba	14.6
146	Ling 3.0 Flash VL	InclusionAI	14.4
146	Nemotron 3.5 Lightning 30B A3B NVFP4	NVIDIA	14.4
148	Qwen3-Omni-30B-A3B-Instruct	Alibaba	14.3
149	Phi-4	Microsoft	14.1
150	GPT-4.1 nano	OpenAI	13.7
151	K-EXAONE 2.0	LG AI Research	13.1
152	Gemma 3 27B	Google	13.0
153	SASarvam 30B	Sarvam	12.6
154	Granite 4.2 8B	IBM	11.2
155	CECeleris-1	Celeris	11.0
156	Exaone 4.0 32B	LG AI Research	10.6
157	Granite 4.2 30B	IBM	10.1
158	LFM2.5-8B-A1B	LiquidAI	9.4
159	Granite 4.2 3B	IBM	9.2
160	Command A+	Cohere	8.9
161	Gemma 4 E4B	Google	8.6
162	Ling 3.0 Tiny	InclusionAI	8.5
163	OPMiniCPM5-2B	OpenBMB	8.4
164	Gemma 4 E2B	Google	6.6
165	Granite-4.0-1B	IBM	6.2
166	LFM2.5-VL-1.6B-Extract	LiquidAI	5.8
167	Granite-4.0-H-1B	IBM	5.2
168	Exaone 4.0 1.2B	LG AI Research	5.0
169	LFM2.5-2.6B	LiquidAI	4.4
170	Granite-4.0-350M	IBM	3.9
171	Granite-4.0-H-350M	IBM	3.8

Evidence key: ObservedLast good

Rows are ordered by the value Artificial Analysis model benchmarks published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.