Benchmarks / Other

AA Openness Index

Artificial Analysis Openness Index (0 to 100): how open a model is, from weight availability and license to transparency about training data and methodology. Not a capability measure.

240models

–Open Index weight

–last updated

Artificial Analysispublisher

#	Model	Effort	Score	Cost / task	Harness
1	K2 Think V2	high	89	–	–
2	K2-V2 (High)	high	89	–	–
3	K2-V2 (Medium)	medium	89	–	–
4	K2-V2 (Low)	low	89	–	–
5	Olmo 3.1 32B Think	–	89	–	–
6	Olmo 3.1 32B Instruct	none	89	–	–
7	Olmo 3 32B Think	–	89	–	–
8	OLMo 2 32B	none	89	–	–
9	Olmo 3 7B Think	–	89	–	–
10	OLMo 2 7B	none	89	–	–
11	Molmo 7B-D	none	89	–	–
12	Olmo 3 7B Instruct	none	89	–	–
13	Apertus 70B Instruct	none	89	–	–
14	Apertus 8B Instruct	none	89	–	–
15	Nemotron 3 Ultra	max	83	–	–
16	Nemotron 3.5 Lightning	–	83	–	–
17	Nemotron 3 Super 120B A12B (Reasoning)	–	83	–	–
18	Nemotron Cascade 2 30B A3B	–	83	–	–
19	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	–	83	–	–
20	MiniCPM5-1B (Reasoning)	–	83	–	–
21	MiniCPM5-1B (Non-reasoning)	none	83	–	–
22	NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)	none	83	–	–
23	NVIDIA Nemotron 3 Nano 4B	–	83	–	–
24	Granite 4.2 30B	–	72	–	–
25	Granite 4.2 8B	–	72	–	–
26	INTELLECT-3 (Based on GLM-4.5-Air)	–	72	–	–
27	Granite 4.2 3B	–	72	–	–
28	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)	–	72	–	–
29	NVIDIA Nemotron Nano 9B V2 (Reasoning)	–	72	–	–
30	NVIDIA Nemotron Nano 9B V2 (Non-reasoning)	none	72	–	–
31	NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)	none	72	–	–
32	Molmo2-8B	none	72	–	–
33	MiniCPM5-2B	–	67	–	–
34	Granite 4.1 30B	none	61	–	–
35	Kimi Linear 48B A3B Instruct	none	61	–	–
36	Granite 4.1 8B	none	61	–	–
37	Granite 4.1 3B	none	61	–	–
38	GLM 4.5	–	56	–	–
39	GLM 4.5 Air	–	56	–	–
40	Nemotron 3 Nano Omni 30B A3B Reasoning	–	56	–	–
41	ERNIE 4.5 300B A47B	none	56	–	–
42	Granite 4.0 H Small	none	56	–	–
43	Granite 4.0 H 1B	none	56	–	–
44	Granite 4.0 Micro	none	56	–	–
45	Granite 4.0 1B	none	56	–	–
46	Granite 4.0 350M	none	56	–	–
47	Granite 4.0 H 350M	none	56	–	–
48	MiMo V2.6 Pro	–	53	–	–
49	MiMo V2.6 Flash	–	53	–	–
50	MiMo V2 Flash	–	53	–	–
51	Llama Nemotron Super 49B v1.5 (Reasoning)	–	53	–	–
52	Llama 3.3	–	53	–	–
53	GLM 4.5 V	–	53	–	–
54	Llama 3.1	max	53	–	–
55	Llama Nemotron Super 49B v1.5 (Non-reasoning)	none	53	–	–
56	DeepSeek V4 Pro 0424	max	50	–	–
57	GLM 5	–	50	–	–
58	DeepSeek V4 Flash 0420	high	50	–	–
59	A.X-K2	–	50	–	–
60	Step 3.5 Flash	–	50	–	–
61	Qwen 3	–	50	–	–
62	Apriel-v1.6-15B-Thinker	–	50	–	–
63	DeepSeek R1	–	50	–	–
64	Magistral Small 1.2	–	50	–	–
65	Phi-4 Mini Instruct	none	50	–	–
66	Phi-4	none	50	–	–
67	Phi-4 Multimodal Instruct	none	50	–	–
68	Gemma 3 27B	none	50	–	–
69	Gemma 3 4B	none	50	–	–
70	Gemma 3n E4B Instruct	none	50	–	–
71	Gemma 3n E2B Instruct	none	50	–	–
72	Gemma 3 1B	none	50	–	–
73	Gemma 3 12B	none	50	–	–
74	Apriel-v1.5-15B-Thinker	–	47	–	–
75	Llama 3.1 70b	–	47	–	–
76	Llama 3.1 405b	–	47	–	–
77	GLM 5.3 Flash	max	44	–	–
78	DeepSeek V4.1 Flash	max	44	–	–
79	DeepSeek V4 Pro 0813	max	44	–	–
80	DeepSeek V4 Flash 0731	max	44	–	–
81	GLM 5.2	max	44	–	–
82	Motif 3	–	44	–	–
83	GLM 5.1	–	44	–	–
84	Hy3	–	44	–	–
85	Ling 3.0 Flash	–	44	–	–
86	Ling 3.0 Flash Fin	–	44	–	–
87	GLM 4.7	–	44	–	–
88	Kimi K2	none	44	–	–
89	Ling 3.0 Flash	–	44	–	–
90	K-EXAONE 2.0 0803	–	44	–	–
91	GLM 4.6	–	44	–	–
92	Muse Glimmer (High)	high	44	–	–
93	DeepSeek V3.2	–	44	–	–
94	GLM 4.7 Flash	–	44	–	–
95	Qwen 3 235B-A22B	–	44	–	–
96	Seed-OSS-36B-Instruct	–	44	–	–
97	Qwen 3 Coder	none	44	–	–
98	LongCat Flash Lite	none	44	–	–
99	Trinity Large Thinking	–	44	–	–
100	Qwen 3 30B-A3B	–	44	–	–
101	Tri-21B-think Preview	–	44	–	–
102	DiffusionGemma 26B A4B	–	44	–	–
103	Ling-1T	none	44	–	–
104	Qwen 3 4B	–	44	–	–
105	Falcon-H1R-7B	–	44	–	–
106	Ling-flash-2.0	none	44	–	–
107	Devstral Small	none	44	–	–
108	MiniCPM-V 4.6 1.3B	none	44	–	–
109	Ling-mini-2.0	none	44	–	–
110	Gemma 3 270M	none	44	–	–
111	North Mini Code	–	42	–	–
112	Nanbeige4.1-3B	–	42	–	–
113	Step3 VL 10B	–	42	–	–
114	Kimi K3	max	39	–	–
115	Qwen 3.8 Flash	–	39	–	–
116	Qwen 3.8 27B	xhigh	39	–	–
117	Nex-N2-Pro	–	39	–	–
118	MiMo V2.5 Pro	–	39	–	–
119	Inkling	xhigh	39	–	–
120	Qwen 3.5 27B	–	39	–	–
121	G9v3-39A5B	–	39	–	–
122	MiMo V2.5	–	39	–	–
123	Qwen 3.5 397B-A17B	none	39	–	–
124	Qwen 3.6 27B	–	39	–	–
125	Step 3.7 Flash	high	39	–	–
126	Qwen 3.5 35B-A3B	–	39	–	–
127	LongCat 2.0	–	39	–	–
128	Qwen 3.6 35B-A3B	–	39	–	–
129	Qwen 3.5 122B-A10B	none	39	–	–
130	Ling-2.6-1T	none	39	–	–
131	Gemma 4 26B	–	39	–	–
132	Ring-2.6-1T	xhigh	39	–	–
133	DeepSeek V3.1	–	39	–	–
134	Gemma 4 31B	–	39	–	–
135	Gemma 4 12B	–	39	–	–
136	Qwen 3.5 9B	none	39	–	–
137	Command A	–	39	–	–
138	Qwen 3.5 4B	–	39	–	–
139	HyperNova 60B 2605 (High, Based on gpt-oss-120b)	high	39	–	–
140	gpt-oss-120b	high	39	–	–
141	Mistral Small 4	–	39	–	–
142	Ling 3.0 Tiny	–	39	–	–
143	Ring-1T	–	39	–	–
144	gpt-oss-20b	low	39	–	–
145	Ling 2.6 Flash	none	39	–	–
146	Mistral Large 3	none	39	–	–
147	Tri-21B-Think	–	39	–	–
148	Gemma 4 E4B (Reasoning)	–	39	–	–
149	Sarvam 105B (High)	high	39	–	–
150	LFM2.5-2.6B	–	39	–	–
151	Gemma 4 E2B (Reasoning)	–	39	–	–
152	Llama 3.3 70b	none	39	–	–
153	Gemma 4 E4B (Non-reasoning)	none	39	–	–
154	Ring-flash-2.0	–	39	–	–
155	Mistral Small 3.2	none	39	–	–
156	Qwen 3.5 2B	–	39	–	–
157	Sarvam 30B (High)	high	39	–	–
158	Gemma 4 E2B (Non-reasoning)	none	39	–	–
159	R1 1776	–	39	–	–
160	Llama 3.2 90b	none	39	–	–
161	Qwen 3.5 0.8B	–	39	–	–
162	Mistral	none	39	–	–
163	Ministral 3 14B	none	39	–	–
164	Jamba Reasoning 3B	–	39	–	–
165	Reka Flash 3	–	39	–	–
166	Ministral 3 8B	none	39	–	–
167	Llama 3.2 11b	none	39	–	–
168	Ministral 3 3B	none	39	–	–
169	Cogito v2.1 (Reasoning)	–	39	–	–
170	Tiny Aya Global	none	36	–	–
171	GLM 5.3	max	33	–	–
172	Step 5 Preview	–	33	–	–
173	MiniMax M3	–	33	–	–
174	Kimi K2.6	–	33	–	–
175	Inkling Small	–	33	–	–
176	Solar Open2 250B	–	33	–	–
177	Kimi K2.5	–	33	–	–
178	Mistral Medium 3.5	high	33	–	–
179	G9v3-3B	–	33	–	–
180	LFM2.5-8B-A1B	–	33	–	–
181	LFM2 2.6B	none	33	–	–
182	Llama 3.1 8b	none	33	–	–
183	LFM2 8B A1B	none	33	–	–
184	LFM2 1.2B	none	33	–	–
185	HyperCLOVA X SEED Think (32B)	–	31	–	–
186	Qwen 3.8 2.4T-A95B	–	28	–	–
187	Muse Spark Open (Xhigh)	xhigh	28	–	–
188	Kimi K2.7 Code	–	28	–	–
189	MiniMax M2.5	–	28	–	–
190	MiniMax M2.1	–	28	–	–
191	MiniMax M2	–	28	–	–
192	K-EXAONE (Reasoning)	–	28	–	–
193	EXAONE 4.5 33B (Reasoning)	–	28	–	–
194	Magistral Medium 1.2	medium	28	–	–
195	K-EXAONE (Non-reasoning)	none	28	–	–
196	Solar Open 100B (Reasoning)	high	28	–	–
197	Llama 4 Maverick	none	28	–	–
198	Devstral 2	none	28	–	–
199	EXAONE 4.0 32B (Reasoning)	–	28	–	–
200	Devstral Small 2	none	28	–	–
201	Llama 4 Scout	none	28	–	–
202	EXAONE 4.0 32B (Non-reasoning)	none	28	–	–
203	LFM2 24B A2B	none	28	–	–
204	Exaone 4.0 1.2B (Reasoning)	–	28	–	–
205	Exaone 4.0 1.2B (Non-reasoning)	none	28	–	–
206	LFM2.5-1.2B-Thinking	–	28	–	–
207	LFM2.5-1.2B-Instruct	none	28	–	–
208	LFM2.5-VL-1.6B	none	28	–	–
209	EXAONE 4.5 33B (Non-reasoning)	none	28	–	–
210	MiniMax M2.7	–	22	–	–
211	Jamba 1.7 Large	none	22	–	–
212	Jamba 1.7 Mini	none	22	–	–
213	Qwen 3 Max	max	17	–	–
214	Opus 4.5	–	11	–	–
215	GPT-5.1	none	11	–	–
216	GPT-5	none	11	–	–
217	Sonnet 4.5	–	11	–	–
218	Grok 4.1 Fast	none	11	–	–
219	Grok 4 Fast	none	11	–	–
220	Doubao Seed Code	–	11	–	–
221	Haiku 4.5	–	11	–	–
222	Gemini 2.5 Flash	none	11	–	–
223	Grok 3 Mini	high	11	–	–
224	Gemini 2.5 Flash Lite	none	11	–	–
225	Mistral Medium 3.1	none	11	–	–
226	Nova Premier	none	11	–	–
227	Devstral Medium	none	11	–	–
228	Solar Pro 2 (Reasoning)	high	11	–	–
229	Solar Pro 2 (Non-reasoning)	none	11	–	–
230	Nova Pro	none	11	–	–
231	Nova Lite	none	11	–	–
232	Nova Micro	none	11	–	–
233	Gemini 3 Pro	high	6	–	–
234	GPT-5 Codex	high	6	–	–
235	Grok 4	–	6	–	–
236	GPT-5 Mini	medium	6	–	–
237	o3	–	6	–	–
238	Gemini 2.5 Pro	–	6	–	–
239	Grok Code Fast 1	–	6	–	–
240	GPT-5 Nano	high	6	–	–