AA-Omniscience

AA-Omniscience Index (-100..100): knowledge accuracy minus hallucination penalty on 6,000 questions across 42 topics.

365models

–Open Index weight

–last updated

Artificial Analysispublisher

#	Model	Effort	Score	Cost / task	Harness
1	Opus 5.5	max	46	$0.161	–
2	GPT-6 Astra	high	44	$0.027	–
3	Fable 5.1	max	43	$0.268	–
4	Fable 5	max	43	$0.070	–
5	Gemini 4	high	42	$0.0096	–
6	GPT-6.1 Sol	max	42	$0.017	–
7	Opus 5	max	37	$0.026	–
8	Sonnet 5.5	max	32	$0.147	–
9	Grok 4.7	xhigh	32	$0.034	–
10	Gemini 3.1 Pro	–	32	$0.018	–
11	Grok 4.6	high	30	$0.0087	–
12	Muse Wonder Full (Max)	max	30	–	–
13	Gemini 3.8 Flash	high	30	$0.013	–
14	Opus 4.8	max	29	$0.025	–
15	Muse Spark 1.1	xhigh	28	$0.014	–
16	Opus 4.7	max	27	–	–
17	Muse Spark 1.2	xhigh	27	$0.010	–
18	GPT-6 Sol	max	27	$0.027	–
19	Gemini 3.7 Flash	high	26	$0.0039	–
20	Muse Wonder	max	25	$0.050	–
21	Grok 4.5	high	25	$0.0065	–
22	Muse Spark 1.3	max	25	$0.023	–
23	Muse Spark Open (Xhigh)	xhigh	25	$0.0080	–
24	Gemini 3.6 Flash	high	22	$0.0052	–
25	GPT-5.6 Sol	max	22	$0.085	–
26	Gemini 3.5 Flash	high	21	$0.014	–
27	JT-4.1 Flash 236B A21B (Reasoning)	–	21	–	–
28	GPT-5.5	xhigh	21	$0.152	–
29	Kimi K3	max	20	$0.133	–
30	Grok 4.3	high	18	$0.0075	–
31	Sonnet 5	max	16	$0.029	–
32	Step 5 Preview	–	16	$0.015	–
33	Gemini 3 Pro	high	15	–	–
34	Grok 4.20	–	15	–	–
35	GLM 5.3	max	14	$0.011	–
36	Opus 4.5	–	14	–	–
37	Opus 4.6	max	14	–	–
38	Qwen 3.7 Max	max	13	$0.014	–
39	Sonnet 4.6	max	12	$0.044	–
40	Qwen 3.8 Max	max	12	$0.0090	–
41	GPT-5.5 Instant	–	11	–	–
42	GPT-5.3 Codex	xhigh	11	–	–
43	Haiku 5.5	max	11	$0.0063	–
44	Motif 3	–	10	–	–
45	Gemini 3 Flash	–	10	–	–
46	Qwen 3.6 Max	max	9	–	–
47	MiMo V2.6 Pro	–	8	$0.0040	–
48	GLM 5.3 Flash	max	7	$0.0006	–
49	Muse Spark	–	7	–	–
50	Grok Build 0.1	–	6	–	–
51	GPT-5.4	xhigh	6	–	–
52	GPT-5.1	high	5	–	–
53	Kimi K2.6	–	5	$0.026	–
54	Gemini 3.5 Flash-Lite	high	5	$0.0031	–
55	MiMo V2 Pro	–	5	–	–
56	GLM 5.2	max	4	$0.0086	–
57	Qwen 3.8 2.4T-A95B	–	4	$0.028	–
58	G9v3-39A5B	–	4	–	–
59	MiMo V2.5 Pro	–	3	$0.0028	–
60	Ling 3.1 Flash	–	2	$0.0040	–
61	Grok 4	–	2	–	–
62	Inkling	xhigh	2	–	–
63	Beam 1.0	–	2	–	–
64	MiniMax M3	–	1	$0.0035	–
65	Qwen 3.7 Plus	–	1	$0.0029	–
66	Qwen 3.6 Plus	–	1	–	–
67	GLM 5.1	–	1	$0.0095	–
68	DeepSeek V4 Pro 0813	max	1	$0.013	–
69	MiniMax M2.7	–	1	–	–
70	GPT-6 Luna	max	1	$0.0025	–
71	GPT-5.2	medium	0	–	–
72	GLM 5	–	0	–	–
73	Solar Pro 4	–	0	$0.0029	–
74	Sonnet 4	–	0	–	–
75	GPT-5.6 Terra	max	0	$0.070	–
76	Sonnet 4.5	–	-0	$0.013	–
77	Nemotron 3 Ultra	max	-0	$0.0031	–
78	Sonnet 3.7	–	-1	–	–
79	MiniCPM5-1B (Non-reasoning)	none	-1	–	–
80	Solar Open2 250B	–	-2	–	–
81	GPT-5.2 Codex	xhigh	-2	–	–
82	Quasar 438B (Max, Based on GLM-5.2)	max	-3	$0.018	–
83	K2 Horizon 375B A23B	–	-3	–	–
84	Command A	–	-4	$0.0000	–
85	Nemotron 3.5 Super VL	–	-4	$0.0033	–
86	Haiku 4.5	–	-4	$0.0045	–
87	G9v3-3B	–	-4	–	–
88	Ling 3.0 Flash	–	-5	–	–
89	DeepSeek V4.1 Flash	max	-5	$0.0098	–
90	Mistral Large 4	–	-5	$0.017	–
91	GPT-5.1 Codex	high	-7	–	–
92	K-EXAONE 2.0 0803	–	-7	–	–
93	Grok 3 Mini	high	-7	–	–
94	Kimi K2.5	–	-7	–	–
95	GPT-5 Codex	high	-8	–	–
96	Qwen 3.8 27B	none	-8	$0.0001	–
97	A.X-K2	–	-8	–	–
98	GPT-5	high	-9	–	–
99	JT-4.1 Flash 236B A21B (Non-reasoning)	none	-9	–	–
100	Inkling Small	–	-9	$0.0060	–
101	Qwen 3.8 Flash	–	-10	$0.0021	–
102	MiMo V2.5	–	-10	$0.0005	–
103	Kimi K2.7 Code	–	-10	$0.021	–
104	GPT-5.6 Luna	max	-10	$0.011	–
105	GPT-4o	none	-11	–	–
106	Agnes 2.5 Pro Beta	–	-11	–	–
107	DeepSeek V4 Pro 0424	high	-11	–	–
108	Solar Mini 4	–	-11	$0.0069	–
109	K2 Horizon MoVA 36B A4B	–	-11	–	–
110	LFM2.5-2.6B	–	-11	–	–
111	GPT-5 Mini	medium	-11	–	–
112	o1	–	-11	–	–
113	MiniCPM5-2B	–	-12	–	–
114	MiMo V2.6 Flash	–	-13	$0.0033	–
115	Granite 4.2 30B	–	-13	$0.0013	–
116	Qwen 3.5	none	-13	–	–
117	DeepSeek V4 Flash 0731	max	-14	$0.020	–
118	Ling 3.0 Flash Fin	–	-15	–	–
119	MiniCPM5-1B (Reasoning)	–	-15	–	–
120	Granite 4.2 3B	–	-15	$0.0006	–
121	MiMo V2	–	-15	–	–
122	o3	–	-16	–	–
123	Gemini 2.5 Pro	–	-16	$0.017	–
124	GPT-5.1 Codex Mini	high	-16	–	–
125	Gemini 3.1 Flash-Lite	high	-16	$0.0016	–
126	K2 Horizon 7B	–	-17	–	–
127	Llama 3.1 405b	none	-17	–	–
128	Granite 4.2 8B	–	-17	$0.0013	–
129	DeepSeek V4 Flash	max	-18	$0.0058	–
130	DeepSeek V3.2	–	-18	–	–
131	Nemotron 3.5 Lightning	–	-18	$0.0006	–
132	Motif 3 (Beta)	–	-18	–	–
133	Ling 3.0 Flash	–	-18	–	–
134	GPT-5.4 Nano	medium	-18	–	–
135	Hy3	–	-18	$0.0035	–
136	GPT-5.4 Mini	xhigh	-19	$0.057	–
137	MiMo V2 Flash	–	-19	–	–
138	Nemotron 3.5 Bolt 74B-A9B	–	-19	–	–
139	GLM 5 V	–	-19	–	–
140	Ling 3.0 Tiny	–	-19	$0.0000	–
141	Gemma 4 E4B (Reasoning)	–	-20	$0.0000	–
142	Qwen 3.6 27B	–	-20	$0.034	–
143	Kimi K2	–	-21	–	–
144	Apodex 1.1	–	-22	$0.015	–
145	Qwen 3.6 35B-A3B	–	-22	$0.0049	–
146	Qwen 3 Coder	none	-22	–	–
147	KAT Coder Pro V2	none	-22	–	–
148	Haiku 3.5	none	-23	–	–
149	DeepSeek V4 Flash 0420	high	-23	$0.0003	–
150	LongCat 2.0	–	-23	$0.024	–
151	Gemma 4 E2B (Reasoning)	–	-24	–	–
152	JT-35B-Flash	none	-24	–	–
153	GPT-5 Nano	medium	-26	–	–
154	Cogito v2.1 (Reasoning)	–	-26	–	–
155	DeepSeek V3.1	–	-26	–	–
156	Magistral Medium 1.2	medium	-27	–	–
157	Magistral Medium 1	medium	-27	–	–
158	GLM 4.6 V	–	-27	–	–
159	Nex-N2-Pro	–	-27	–	–
160	GLM 4.5	–	-27	–	–
161	DeepSeek R1	–	-27	–	–
162	K2 Horizon 3.7B	–	-29	–	–
163	Gemma 3 270M	none	-29	–	–
164	Gemini 2.5 Flash	–	-30	–	–
165	Grok 4.1 Fast	–	-30	–	–
166	Grok 4 Fast	–	-30	–	–
167	Mistral Small 4	–	-30	$0.0004	–
168	Qwen 3.5 397B-A17B	–	-31	$0.015	–
169	Llama 3.1 8b	none	-31	–	–
170	Mistral Medium 3	none	-31	–	–
171	Devstral Medium	none	-32	–	–
172	GLM 4.6	none	-32	–	–
173	Muse Glimmer (High)	high	-33	$0.0025	–
174	MiniMax M2.1	–	-33	–	–
175	Grok 3	none	-33	–	–
176	LFM2.5-8B-A1B	–	-33	–	–
177	GPT-4	none	-34	–	–
178	Mistral Large 2	none	-34	–	–
179	ERNIE 4.5 300B A47B	none	-35	–	–
180	Doubao Seed Code	–	-35	–	–
181	Nova Premier	none	-35	–	–
182	o4 mini	high	-36	–	–
183	GLM 4.7	–	-36	–	–
184	Qwen 3 Max	max	-37	–	–
185	Mistral Medium 3.5	high	-37	$0.014	–
186	Grok Code Fast 1	–	-37	–	–
187	Step 3.7 Flash	high	-37	–	–
188	Ring-2.6-1T	xhigh	-38	$0.0049	–
189	KAT-Coder-Pro V1	none	-38	–	–
190	MiniMax M2.5	–	-39	–	–
191	Mercury 2.5	–	-39	$0.0015	–
192	Mistral Large 3	none	-40	$0.0001	–
193	GPT-4.1	none	-40	–	–
194	K2 Think V2	high	-40	–	–
195	DeepSeek V3	none	-41	$0.0000	–
196	Llama 3.1	none	-41	–	–
197	Gemma 4 E4B (Non-reasoning)	none	-41	–	–
198	Qwen 3.5 122B-A10B	–	-42	$0.014	–
199	Nemotron 3 Super 120B A12B (Reasoning)	–	-42	$0.0021	–
200	NVIDIA Nemotron Nano 9B V2 (Reasoning)	–	-42	–	–
201	Step 3.5 Flash	–	-42	–	–
202	Llama 4 Maverick	none	-42	$0.0001	–
203	Nova Lite	none	-42	–	–
204	Nanbeige4.1-3B	–	-43	–	–
205	Gemini 2.0 Flash	none	-43	–	–
206	o3 mini	high	-43	–	–
207	Gemini 2.5 Flash Lite	none	-43	–	–
208	Llama 3.1 70b	none	-43	–	–
209	Olmo 3.1 32B Think	–	-43	–	–
210	Qwen 3.5 27B	–	-44	–	–
211	Qwen 3 235B-A22B	none	-44	–	–
212	Trinity Large Thinking	–	-44	$0.0050	–
213	Step 3.5 Flash 2603	–	-44	–	–
214	Magistral Small 1	–	-44	–	–
215	ERNIE 5.0 Thinking Preview	–	-46	–	–
216	Llama Nemotron Super 49B v1.5 (Non-reasoning)	none	-46	–	–
217	Mistral Medium 3.1	none	-46	–	–
218	GLM 4.5 V	–	-46	–	–
219	Nova 2.0 Pro Preview (Low)	low	-47	–	–
220	MiniMax M2	–	-47	–	–
221	Qwen 3	–	-47	–	–
222	Devstral 2	none	-47	–	–
223	K2-V2 (Low)	low	-47	–	–
224	Llama Nemotron Super 49B v1.5 (Reasoning)	–	-48	–	–
225	Nova Pro	none	-48	–	–
226	Gemma 4 31B	–	-48	$0.0000	–
227	Qwen 3.5 35B-A3B	–	-48	–	–
228	MiniMax M1	–	-49	–	–
229	Nova Micro	none	-49	–	–
230	North Mini Code	–	-49	$0.0000	–
231	Haiku 3	none	-49	–	–
232	Qwen 3 14B	–	-49	–	–
233	K2-V2 (Medium)	medium	-49	–	–
234	Nova 2.0 Pro Preview (Medium)	medium	-49	–	–
235	gpt-oss-120b	high	-49	$0.0011	–
236	Nova 2.0 Pro Preview (Non-reasoning)	none	-50	–	–
237	Llama 3.3	none	-50	–	–
238	Nova 2.0 Omni (Low)	low	-50	–	–
239	Qwen 3 32B	–	-50	–	–
240	Mistral Small 3.2	none	-51	$0.0000	–
241	Ling-2.6-1T	none	-51	–	–
242	EXAONE 4.5 33B (Reasoning)	–	-51	–	–
243	Mercury 2	high	-51	$0.0035	–
244	Mistral Small 3.1	none	-51	$0.0000	–
245	Gemma 4 26B	–	-51	–	–
246	Devstral Small	none	-51	–	–
247	INTELLECT-3 (Based on GLM-4.5-Air)	–	-51	–	–
248	Ring-1T	–	-51	–	–
249	Nemotron Cascade 2 30B A3B	–	-51	–	–
250	Qwen 3 30B-A3B	–	-51	–	–
251	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	–	-52	$0.0010	–
252	Llama 4 Scout	none	-52	$0.0000	–
253	Olmo 3.1 32B Instruct	none	-52	–	–
254	Seed-OSS-36B-Instruct	–	-52	–	–
255	Gemma 4 12B	–	-53	–	–
256	Qwen 2.5 72B	none	-53	–	–
257	HyperCLOVA X SEED Think (32B)	–	-53	–	–
258	Solar Pro 3	high	-53	$0.0013	–
259	Qwen 3.5 9B	–	-53	$0.0028	–
260	GPT-4.1 Mini	none	-54	–	–
261	LFM2 2.6B	none	-54	–	–
262	Nova 2.0 Lite (Low)	low	-54	–	–
263	Llama 3.3 70b	none	-54	–	–
264	Llama 3 70b	none	-54	–	–
265	Solar Open 100B (Reasoning)	high	-55	–	–
266	Qwen 3.5 0.8B	–	-55	–	–
267	HyperNova 60B 2605 (High, Based on gpt-oss-120b)	high	-55	–	–
268	Llama 3.2 1b	none	-55	–	–
269	Apriel-v1.5-15B-Thinker	–	-55	–	–
270	Jamba 1.7 Large	none	-55	–	–
271	Nova 2.0 Lite (High)	high	-55	–	–
272	Ling-1T	none	-55	–	–
273	K2-V2 (High)	high	-56	–	–
274	Phi-4	none	-56	–	–
275	Solar Pro 2 (Reasoning)	high	-56	–	–
276	Nova 2.0 Lite (Medium)	medium	-57	–	–
277	NVIDIA Nemotron Nano 9B V2 (Non-reasoning)	none	-57	–	–
278	Nemotron 3 Nano Omni 30B A3B Reasoning	–	-57	–	–
279	GPT-4.1 Nano	none	-58	–	–
280	Devstral Small 2	none	-58	–	–
281	Jamba Reasoning 3B	–	-58	–	–
282	K-EXAONE (Reasoning)	–	-58	–	–
283	Olmo 3 32B Think	–	-58	–	–
284	LFM2 24B A2B	none	-58	–	–
285	Qwen 3.5 4B	–	-58	–	–
286	gpt-oss-20b	low	-59	–	–
287	Ring-flash-2.0	–	-59	–	–
288	Step3 VL 10B	–	-59	–	–
289	Nova 2.0 Omni (Medium)	medium	-59	–	–
290	Mi:dm K 2.5 Pro	–	-59	–	–
291	DiffusionGemma 26B A4B	–	-59	–	–
292	Apriel-v1.6-15B-Thinker	–	-59	–	–
293	Sarvam 105B (High)	high	-59	–	–
294	Nova 2.0 Lite (Non-reasoning)	none	-60	–	–
295	EXAONE 4.0 32B (Reasoning)	–	-60	–	–
296	Qwen 3.5 2B	–	-60	–	–
297	Gemma 4 E2B (Non-reasoning)	none	-60	–	–
298	Qwen 3 4B	–	-60	–	–
299	Granite 4.0 H Small	none	-61	–	–
300	Phi-4 Mini Instruct	none	-61	–	–
301	Falcon-H1R-7B	–	-61	–	–
302	GLM 4.5 Air	–	-62	–	–
303	Apertus 70B Instruct	none	-62	–	–
304	Sarvam M (Reasoning, Based on Mistral Small 3.1)	high	-62	–	–
305	Motif-2-12.7B-Reasoning	–	-62	–	–
306	Solar Pro 2 (Non-reasoning)	none	-62	–	–
307	Tri-21B-think Preview	–	-62	–	–
308	Tri-21B-Think	–	-63	–	–
309	Llama 3.2 11b	none	-63	–	–
310	GLM 4.7 Flash	–	-63	–	–
311	Ling-flash-2.0	none	-63	–	–
312	Mi:dm K 2.5 Pro Preview	–	-63	–	–
313	EXAONE 4.0 32B (Non-reasoning)	none	-63	–	–
314	Ministral 3 3B	none	-64	$0.0000	–
315	JT-MINI	none	-64	–	–
316	Granite 4.1 8B	none	-64	–	–
317	Nova 2.0 Omni (Non-reasoning)	none	-64	–	–
318	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)	–	-64	–	–
319	Qwen 3 8B	–	-65	–	–
320	Mistral	none	-65	–	–
321	Magistral Small 1.2	–	-65	–	–
322	Reka Flash 3	–	-65	–	–
323	K-EXAONE (Non-reasoning)	none	-65	–	–
324	Ling 2.6 Flash	none	-66	–	–
325	Ministral 3 14B	none	-66	$0.0000	–
326	Gemma 3 27B	none	-67	$0.0000	–
327	Granite 4.1 30B	none	-68	–	–
328	Ministral 3 8B	none	-69	$0.0000	–
329	Molmo2-8B	none	-69	–	–
330	NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)	none	-69	–	–
331	Granite 4.0 350M	none	-69	–	–
332	LongCat Flash Lite	none	-70	–	–
333	Llama 3 8b	none	-70	–	–
334	NVIDIA Nemotron 3 Nano 4B	–	-71	–	–
335	Sarvam 30B (High)	high	-72	–	–
336	Celeris-1	none	-72	$0.0000	–
337	Olmo 3 7B Think	–	-72	–	–
338	LFM2.5-1.2B-Instruct	none	-72	–	–
339	NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)	none	-72	–	–
340	Granite 4.0 H 1B	none	-72	–	–
341	LFM2 1.2B	none	-72	–	–
342	K2 Horizon 0.9B	–	-73	–	–
343	Jamba 1.7 Mini	none	-73	–	–
344	Gemma 3 1B	none	-75	–	–
345	LFM2 8B A1B	none	-77	–	–
346	Granite 4.1 3B	none	-77	–	–
347	Olmo 3 7B Instruct	none	-77	–	–
348	Gemma 3 12B	none	-77	$0.0000	–
349	Apertus 8B Instruct	none	-77	–	–
350	Granite 3.3 8B (Non-reasoning)	none	-77	–	–
351	Qwen 3 1.7B	–	-78	–	–
352	Granite 4.0 Micro	none	-78	–	–
353	Ling-mini-2.0	none	-78	–	–
354	Qwen 3 0.6B	–	-79	–	–
355	Exaone 4.0 1.2B (Reasoning)	–	-80	–	–
356	LFM2.5-1.2B-Thinking	–	-80	–	–
357	Gemma 3n E2B Instruct	none	-80	–	–
358	Granite 4.0 H 350M	none	-81	–	–
359	Gemma 3n E4B Instruct	none	-81	–	–
360	Granite 4.0 1B	none	-82	–	–
361	Exaone 4.0 1.2B (Non-reasoning)	none	-82	–	–
362	MiniCPM-V 4.6 1.3B	none	-83	–	–
363	Gemma 3 4B	none	-83	–	–
364	Tiny Aya Global	none	-84	–	–
365	LFM2.5-VL-1.6B	none	-84	–	–