Benchmarks / Composite

AA Intelligence Index v4.3.2

Artificial Analysis' composite intelligence score (v4.3.2), combining 10 evaluations: AA-Briefcase v1.1, GDPval-AA v2.1, AutomationBench-AA, Terminal-Bench 4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, AA-LCR v1.1.

441models

–Open Index weight

–last updated

Artificial Analysispublisher

#	Model	Effort	Score	Cost / task	Harness
1	Opus 5.5	max	58	$5.98	–
2	Sonnet 5.5	max	56	$5.46	–
3	Muse Wonder Full (Max)	max	56	–	–
4	Fable 5.1	max	53	$7.63	–
5	GPT-6 Astra	max	53	$3.26	–
6	Gemini 4	high	53	$1.99	–
7	Muse Wonder	max	52	$4.52	–
8	GPT-6.1 Sol	max	52	$0.724	–
9	Opus 5	max	51	$5.86	–
10	Fable 5	max	50	$8.75	–
11	Muse Spark 1.3	max	48	$1.60	–
12	GPT-6 Sol	max	48	$1.04	–
13	GPT-5.6 Sol	max	47	$1.99	–
14	Grok 4.7	xhigh	46	$3.74	–
15	MiMo V2.6 Pro	–	46	$0.133	–
16	Qwen 3.8 Max	max	45	$5.41	–
17	GLM 5.3	max	45	$2.01	–
18	Grok 4.6	high	44	$1.48	–
19	Step 5 Preview	–	44	$1.03	–
20	Kimi K3	max	44	$2.00	–
21	Haiku 5.5	max	43	$0.213	–
22	GPT-5.6 Terra	max	42	$1.40	–
23	GLM 5.3 Flash	max	42	$0.284	–
24	Opus 4.8	max	42	$4.08	–
25	Ling 3.1 Flash	–	41	$1.000	–
26	Gemini 3.8 Flash	high	41	$1.24	–
27	Opus 4.7	max	41	–	–
28	Qwen 3.8 2.4T-A95B	–	40	$2.16	–
29	Qwen 3.8 Flash	–	40	$0.372	–
30	Gemini 3.7 Flash	medium	40	–	–
31	Muse Spark 1.2	xhigh	40	$0.975	–
32	DeepSeek V4.1 Flash	max	39	$0.265	–
33	GPT-5.4	xhigh	39	–	–
34	Muse Spark Open (Xhigh)	xhigh	39	$0.844	–
35	Grok 4.5	high	39	$0.881	–
36	Mistral Large 4	–	38	$1.13	–
37	GPT-5.5	xhigh	38	$2.63	–
38	Sonnet 5	max	38	$5.09	–
39	GPT-6 Luna	max	38	$0.068	–
40	MiMo V2.6 Flash	–	38	$0.062	–
41	GPT-5.6 Luna	max	37	$0.178	–
42	DeepSeek V4 Pro 0813	max	36	$0.674	–
43	Agnes 2.5 Pro Beta	–	35	–	–
44	DeepSeek V4 Flash	max	35	$0.314	–
45	DeepSeek V4 Flash 0731	max	34	$0.220	–
46	Gemini 3.6 Flash	high	34	$1.60	–
47	JT-4.1 Flash 236B A21B (Reasoning)	–	34	–	–
48	Muse Spark 1.1	xhigh	34	$1.38	–
49	GLM 5.2	max	34	$1.47	–
50	Qwen 3.8 27B	xhigh	34	$1.01	–
51	Gemini 3.5 Flash	medium	34	–	–
52	Motif 3	–	34	–	–
53	GPT-5.3 Codex	xhigh	33	–	–
54	Motif 3 (Beta)	–	32	–	–
55	Opus 4.6	max	32	–	–
56	Muse Spark	–	31	–	–
57	K2 Horizon 375B A23B	–	31	–	–
58	DeepSeek V4 Pro 0424	max	30	$0.122	–
59	GPT-5.2	xhigh	30	–	–
60	Beam 1.0	–	30	–	–
61	Sonnet 4.6	max	30	$2.49	–
62	Gemini 3.1 Pro	–	30	$1.30	–
63	Qwen 3.7 Max	max	29	$1.15	–
64	MiniMax M3	–	29	$0.508	–
65	Opus 4.5	–	29	–	–
66	MiMo V2 Pro	–	29	–	–
67	GPT-5.2 Codex	xhigh	29	–	–
68	Qwen 3.6 Max	max	28	–	–
69	Nex-N2-Pro	–	28	–	–
70	Gemini 3 Pro	high	28	–	–
71	GLM 5	–	28	–	–
72	JT-4.1 Flash 236B A21B (Non-reasoning)	none	27	–	–
73	Grok Build 0.1	–	27	–	–
74	Qwen 3.6 Plus	–	27	–	–
75	Kimi K2.6	–	27	$0.800	–
76	Quasar 438B (Max, Based on GLM-5.2)	max	27	$2.02	–
77	Solar Pro 4	–	27	$0.711	–
78	Nemotron 3.5 Super VL	–	26	$0.251	–
79	Apodex 1.1	–	26	$0.464	–
80	Gemini 3 Flash	–	26	–	–
81	GLM 5.1	–	26	$1.02	–
82	GPT-5.5 Instant	–	26	$0.691	–
83	MiMo V2.5 Pro	–	26	$0.054	–
84	Kimi K2.7 Code	–	26	$0.541	–
85	Inkling Small	–	26	$0.089	–
86	Grok 4.20	–	26	–	–
87	K2 Horizon MoVA 36B A4B	–	25	–	–
88	Hy3	–	25	$0.072	–
89	Qwen 3.7 Plus	–	25	$0.217	–
90	MiMo V2	–	25	–	–
91	Inkling	xhigh	25	–	–
92	GPT-5 Codex	high	25	–	–
93	Grok 4.3	high	25	$0.205	–
94	Solar Open2 250B	–	25	–	–
95	GPT-5.1	high	25	–	–
96	Ling 3.0 Flash	–	25	–	–
97	DeepSeek V4 Flash 0420	high	24	$0.103	–
98	GPT-5.4 Mini	xhigh	24	$0.447	–
99	Solar Mini 4	–	24	$0.366	–
100	GPT-5.1 Codex	high	24	–	–
101	GLM 5 V	–	23	–	–
102	Kimi K2.5	–	23	–	–
103	GPT-5	high	23	–	–
104	Nemotron 3 Ultra	max	23	$0.598	–
105	Qwen 3.5 27B	–	23	–	–
106	Opus 4.1	–	23	–	–
107	MiniMax M2.5	–	23	–	–
108	MiniMax M2.7	–	23	–	–
109	Ling 3.0 Flash Fin	–	23	–	–
110	Grok 4	–	22	–	–
111	MiMo V2 Flash	–	22	–	–
112	GLM 4.7	–	22	–	–
113	Gemini 3.5 Flash-Lite	high	22	$0.193	–
114	Kimi K2	–	22	–	–
115	o3 pro	–	22	–	–
116	G9v3-39A5B	–	22	–	–
117	MiMo V2.5	–	22	$0.019	–
118	KAT Coder Pro V2	none	22	–	–
119	DeepSeek V3.2	–	21	–	–
120	A.X-K2	–	21	–	–
121	Qwen 3.5 397B-A17B	none	21	–	–
122	Qwen 3.6 27B	–	21	$0.621	–
123	Nemotron 3.5 Bolt 74B-A9B	–	21	–	–
124	Qwen 3 Max	max	21	–	–
125	MiniMax M2.1	–	21	–	–
126	GPT-5.4 Nano	xhigh	21	$0.183	–
127	Sonnet 4.5	–	21	$0.557	–
128	Opus 4	–	21	–	–
129	GPT-5 Mini	medium	21	–	–
130	K2 Horizon 7B	–	21	–	–
131	Qwen 3.5	none	20	–	–
132	GPT-5.1 Codex Mini	high	20	–	–
133	Grok 4.1 Fast	–	20	–	–
134	o3	–	20	–	–
135	Ling 3.0 Flash	–	20	–	–
136	K-EXAONE 2.0 0803	–	20	–	–
137	Step 3.7 Flash	high	19	–	–
138	Qwen 3.5 35B-A3B	–	19	–	–
139	LongCat 2.0	–	19	$0.146	–
140	Sonnet 4	–	19	–	–
141	JT-35B-Flash	none	19	–	–
142	MiniMax M2	–	19	–	–
143	KAT-Coder-Pro V1	none	19	–	–
144	GLM 4.6	–	19	–	–
145	Qwen 3.6 35B-A3B	–	18	$0.475	–
146	Grok 4 Fast	–	18	–	–
147	Qwen 3.5 122B-A10B	none	18	–	–
148	Sonnet 3.7	–	18	–	–
149	Muse Glimmer (High)	high	17	$0.054	–
150	Ling-2.6-1T	none	17	–	–
151	Step 3.5 Flash 2603	–	17	–	–
152	Doubao Seed Code	–	17	–	–
153	Haiku 4.5	–	17	$0.277	–
154	Gemma 4 26B	–	17	–	–
155	o4 mini	high	17	–	–
156	Ring-2.6-1T	xhigh	17	$0.287	–
157	Step 3.5 Flash	–	17	–	–
158	Gemini 2.5 Pro	–	16	$0.330	–
159	K2 Horizon 3.7B	–	16	–	–
160	Gemini 3.1 Flash-Lite	high	16	$0.060	–
161	Gemini 2.5 Flash	–	15	–	–
162	o1	–	15	–	–
163	GLM 4.7 Flash	–	15	–	–
164	DeepSeek V3.1	–	15	–	–
165	Gemma 4 31B	–	15	$0.0000	–
166	Grok 3 Mini	high	15	–	–
167	K-EXAONE (Reasoning)	–	14	–	–
168	ERNIE 5.0 Thinking Preview	–	14	–	–
169	Mistral Medium 3.5	high	14	$0.502	–
170	Gemma 4 12B	–	14	–	–
171	Nova 2.0 Pro Preview (Medium)	medium	14	–	–
172	Grok Code Fast 1	–	14	–	–
173	Apriel-v1.5-15B-Thinker	–	14	–	–
174	Nova 2.0 Omni (Medium)	medium	14	–	–
175	Qwen 3	–	13	–	–
176	Apriel-v1.6-15B-Thinker	–	13	–	–
177	Nova 2.0 Lite (High)	high	13	–	–
178	Qwen 3.5 9B	none	13	–	–
179	EXAONE 4.5 33B (Reasoning)	–	13	–	–
180	Command A	–	13	$0.0000	–
181	Qwen 3.5 4B	–	13	–	–
182	DeepSeek R1	–	13	–	–
183	GPT-5 Nano	high	13	–	–
184	Nemotron 3.5 Lightning	–	13	$0.093	–
185	Nemotron 3 Super 120B A12B (Reasoning)	–	13	$1.64	–
186	Nova 2.0 Pro Preview (Low)	low	13	–	–
187	GLM 4.5	–	13	–	–
188	Qwen 3 235B-A22B	–	13	–	–
189	GPT-4.1	none	13	–	–
190	Granite 4.2 30B	–	13	$0.072	–
191	Nova 2.0 Lite (Medium)	medium	12	–	–
192	o3 mini	–	12	–	–
193	MiniCPM5-2B	–	12	–	–
194	o1 pro	–	12	–	–
195	Mercury 2.5	–	12	$0.117	–
196	JT-MINI	none	12	–	–
197	Grok 3	none	12	–	–
198	Seed-OSS-36B-Instruct	–	12	–	–
199	Qwen 3 Coder	none	12	–	–
200	Magistral Medium 1.2	medium	12	–	–
201	Sonar Reasoning Pro	–	12	–	–
202	Nova 2.0 Lite (Low)	low	12	–	–
203	HyperNova 60B 2605 (High, Based on gpt-oss-120b)	high	12	–	–
204	MiniMax M1	–	12	–	–
205	Nemotron Cascade 2 30B A3B	–	12	–	–
206	gpt-oss-120b	high	12	$0.107	–
207	K2 Think V2	high	11	–	–
208	LongCat Flash Lite	none	11	–	–
209	HyperCLOVA X SEED Think (32B)	–	11	–	–
210	Mistral Small 4	–	11	$0.015	–
211	GLM 4.6 V	–	11	–	–
212	K-EXAONE (Non-reasoning)	none	11	–	–
213	Nova 2.0 Omni (Low)	low	11	–	–
214	GLM 4.5 Air	–	11	–	–
215	Granite 4.2 8B	–	11	$0.024	–
216	Ling 3.0 Tiny	–	11	$0.0000	–
217	Mi:dm K 2.5 Pro	–	11	–	–
218	Ring-1T	–	11	–	–
219	G9v3-3B	–	11	–	–
220	Trinity Large Thinking	–	11	$0.119	–
221	Mercury 2	high	11	$0.039	–
222	INTELLECT-3 (Based on GLM-4.5-Air)	–	11	–	–
223	Solar Open 100B (Reasoning)	high	10	–	–
224	Gemini 2.5 Flash Lite	–	10	–	–
225	Nemotron 3 Nano Omni 30B A3B Reasoning	–	10	–	–
226	GPT-4.1 Mini	none	10	–	–
227	Nova 2.0 Pro Preview (Non-reasoning)	none	10	–	–
228	gpt-oss-20b	low	10	–	–
229	North Mini Code	–	10	$0.0000	–
230	K2-V2 (High)	high	10	–	–
231	Qwen 3 30B-A3B	–	10	$0.066	–
232	o1 mini	–	10	–	–
233	DeepSeek V3	none	10	$0.021	–
234	Ling 2.6 Flash	none	10	–	–
235	Tri-21B-think Preview	–	10	–	–
236	GPT-4.5	none	10	–	–
237	DiffusionGemma 26B A4B	–	10	–	–
238	QwQ 32B	–	9	–	–
239	Gemini 2.0 Flash	–	9	–	–
240	Llama 4 Maverick	none	9	$0.038	–
241	Mistral Large 3	none	9	$0.031	–
242	Motif-2-12.7B-Reasoning	–	9	–	–
243	Mistral Medium 3.1	none	9	–	–
244	Ling-1T	none	9	–	–
245	Nova Premier	none	9	–	–
246	Solar Pro 2 (Preview) (Reasoning)	high	9	–	–
247	Granite 4.2 3B	–	9	$0.0060	–
248	Magistral Medium 1	medium	9	–	–
249	Mistral Medium 3	none	9	–	–
250	K2-V2 (Medium)	medium	9	–	–
251	Llama Nemotron Super 49B v1.5 (Reasoning)	–	9	–	–
252	Devstral Medium	none	9	–	–
253	Tri-21B-Think	–	9	–	–
254	GPT-4o	none	9	–	–
255	Haiku 3.5	none	9	–	–
256	Llama 3.3	–	9	–	–
257	Gemma 4 E4B (Reasoning)	–	9	$0.0082	–
258	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	–	9	$0.017	–
259	Qwen 3 4B	–	9	–	–
260	MiniCPM5-1B (Reasoning)	–	9	–	–
261	Sarvam 105B (High)	high	9	–	–
262	Gemini 2.0 Pro	none	9	–	–
263	Nova 2.0 Lite (Non-reasoning)	none	9	–	–
264	Devstral Small	none	9	–	–
265	Opus 3	none	9	–	–
266	MiniCPM5-1B (Non-reasoning)	none	9	–	–
267	Sonar Reasoning	–	9	–	–
268	Devstral 2	none	9	–	–
269	Magistral Small 1.2	–	9	–	–
270	Nanbeige4.1-3B	–	8	–	–
271	LFM2.5-2.6B	–	8	–	–
272	Magistral Small 1	–	8	–	–
273	EXAONE 4.0 32B (Reasoning)	–	8	–	–
274	Nova 2.0 Omni (Non-reasoning)	none	8	–	–
275	Qwen 2.5 Max	none	8	–	–
276	Llama 3.1 70b	–	8	–	–
277	Gemini 1.5 Pro	none	8	–	–
278	Solar Pro 2 (Preview) (Non-reasoning)	none	8	–	–
279	Sonnet 3.5	none	8	–	–
280	Falcon-H1R-7B	–	8	–	–
281	GPT-4.1 Nano	none	8	–	–
282	Solar Pro 3	high	8	$0.079	–
283	Ling-flash-2.0	none	8	–	–
284	Gemma 4 E2B (Reasoning)	–	8	–	–
285	Qwen 2.5 72B	none	8	–	–
286	Sonar	none	8	–	–
287	Step3 VL 10B	–	8	–	–
288	Llama 3.3 70b	none	8	–	–
289	Sonar Pro	none	8	–	–
290	QwQ 32B-Preview	–	8	–	–
291	GLM 4.5 V	–	8	–	–
292	Mistral Large 2	none	8	–	–
293	Devstral Small 2	none	8	–	–
294	Llama 3.1	max	8	–	–
295	ERNIE 4.5 300B A47B	none	8	–	–
296	Llama 3.1 405b	–	7	–	–
297	Solar Pro 2 (Reasoning)	high	7	–	–
298	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)	–	7	–	–
299	Gemma 4 E4B (Non-reasoning)	none	7	–	–
300	Granite 4.1 30B	none	7	–	–
301	NVIDIA Nemotron Nano 9B V2 (Reasoning)	–	7	–	–
302	Gemini 2.0 Flash-Lite	none	7	–	–
303	Llama Nemotron Super 49B v1.5 (Non-reasoning)	none	7	–	–
304	Qwen 3 32B	none	7	–	–
305	K2-V2 (Low)	low	7	–	–
306	Kimi Linear 48B A3B Instruct	none	7	–	–
307	LFM2.5-8B-A1B	–	7	–	–
308	Ring-flash-2.0	–	7	–	–
309	Pixtral Large	none	7	–	–
310	Olmo 3.1 32B Think	–	7	–	–
311	Mistral Small 3.1	none	7	$0.026	–
312	Grok 2	none	7	–	–
313	Gemini 1.5 Flash	none	7	–	–
314	GPT-4	none	7	–	–
315	Solar Pro 2 (Non-reasoning)	none	7	–	–
316	Mistral Small 3.2	none	7	$0.109	–
317	Nova Pro	none	7	–	–
318	Qwen 3.5 2B	–	7	–	–
319	Llama 3.1 8b	none	7	–	–
320	Grok Beta	none	7	–	–
321	Qwen 2.5 32B	none	7	–	–
322	NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)	none	7	–	–
323	NVIDIA Nemotron Nano 9B V2 (Non-reasoning)	none	7	–	–
324	Qwen 2.5 Coder	none	7	–	–
325	Qwen 3 14B	none	7	–	–
326	Mistral Small 3	none	7	–	–
327	Nova Lite	none	7	–	–
328	GPT-4o mini	none	7	–	–
329	DeepSeek V2.5	none	7	–	–
330	Granite 4.1 8B	none	7	–	–
331	Sarvam 30B (High)	high	7	–	–
332	Olmo 3.1 32B Instruct	none	6	–	–
333	Mistral	none	6	–	–
334	Gemma 4 E2B (Non-reasoning)	none	6	–	–
335	Olmo 3 32B Think	–	6	–	–
336	Llama 4 Scout	none	6	$0.112	–
337	R1 1776	–	6	–	–
338	Qwen 2.5 Turbo	none	6	–	–
339	Reka Flash (Sep '24)	none	6	–	–
340	Llama 3.2 90b	none	6	–	–
341	Solar Mini	none	6	–	–
342	Celeris-1	none	6	$0.050	–
343	Grok 1	none	6	–	–
344	Qwen 2 72B	none	6	–	–
345	Phi-4 Mini Instruct	none	6	–	–
346	EXAONE 4.0 32B (Non-reasoning)	none	6	–	–
347	Qwen 3.5 0.8B	–	6	–	–
348	Jamba 1.7 Large	none	6	–	–
349	Granite 4.0 H Small	none	6	–	–
350	Ministral 3 14B	none	6	$0.019	–
351	Jamba 1.5 Large	none	6	–	–
352	Qwen 3 8B	none	6	–	–
353	DeepSeek-Coder-V2	none	6	–	–
354	OLMo 2 32B	none	6	–	–
355	Jamba 1.6 Large	none	6	–	–
356	LFM2 24B A2B	none	6	–	–
357	Phi-4	none	6	–	–
358	Sonnet 3	none	6	–	–
359	Nova Micro	none	6	–	–
360	Granite 4.1 3B	none	6	–	–
361	Mistral Small	none	6	–	–
362	Gemini 1.0 Ultra	none	6	–	–
363	Phi-3 Mini Instruct 3.8B	none	6	–	–
364	NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)	none	6	–	–
365	Gemma 3n E4B Instruct Preview (May '25)	none	6	–	–
366	Phi-4 Multimodal Instruct	none	6	–	–
367	Mistral Large	none	6	–	–
368	Mixtral 8x22B Instruct	none	6	–	–
369	Llama 2 7b	none	6	–	–
370	Llama 3.2 3b	none	6	–	–
371	MiniCPM-V 4.6 1.3B	none	6	–	–
372	Jamba Reasoning 3B	–	6	–	–
373	Qwen 1.5 110B	none	6	–	–
374	Reka Flash 3	–	6	–	–
375	Olmo 3 7B Think	–	6	–	–
376	Claude 2.1	none	6	–	–
377	Haiku 3	none	6	–	–
378	OLMo 2 7B	none	6	–	–
379	Molmo 7B-D	none	6	–	–
380	Ling-mini-2.0	none	6	–	–
381	DeepSeek V2	none	6	–	–
382	Claude 2.0	none	6	–	–
383	Mistral Medium	none	5	–	–
384	GPT-3.5	none	5	–	–
385	Ministral 3 8B	none	5	$0.011	–
386	Llama 3 70b	none	5	–	–
387	Qwen Chat 72B	none	5	–	–
388	Arctic Instruct	none	5	–	–
389	LFM 40B	none	5	–	–
390	Llama 3.2 11b	none	5	–	–
391	PALM-2	none	5	–	–
392	Gemini 1.0 Pro	none	5	–	–
393	DeepSeek Coder V2 Lite Instruct	none	5	–	–
394	Sarvam M (Reasoning, Based on Mistral Small 3.1)	high	5	–	–
395	Llama 2 70b	none	5	–	–
396	DeepSeek LLM 67B Chat (V1)	none	5	–	–
397	Llama 2 13b	none	5	–	–
398	Command-R+ (Apr '24)	none	5	–	–
399	OpenChat 3.5 (1210)	none	5	–	–
400	DBRX Instruct	none	5	–	–
401	Exaone 4.0 1.2B (Reasoning)	–	5	–	–
402	Olmo 3 7B Instruct	none	5	–	–
403	Exaone 4.0 1.2B (Non-reasoning)	none	5	–	–
404	LFM2.5-1.2B-Thinking	–	5	–	–
405	Jamba 1.7 Mini	none	5	–	–
406	LFM2 2.6B	none	5	–	–
407	LFM2.5-1.2B-Instruct	none	5	–	–
408	Jamba 1.5 Mini	none	5	–	–
409	Granite 4.0 H 1B	none	5	–	–
410	Qwen 3 1.7B	–	5	–	–
411	Jamba 1.6 Mini	none	5	–	–
412	Mixtral 8x7B Instruct	none	5	–	–
413	Gemma 3 270M	none	5	–	–
414	Apertus 70B Instruct	none	5	–	–
415	Granite 4.0 Micro	none	5	–	–
416	Claude Instant	none	5	–	–
417	Command-R (Mar '24)	none	5	–	–
418	Qwen Chat 14B	none	5	–	–
419	Llama 6 5b	none	5	–	–
420	Granite 4.0 1B	none	5	–	–
421	Molmo2-8B	none	5	–	–
422	LFM2 8B A1B	none	5	–	–
423	Granite 3.3 8B (Non-reasoning)	none	5	–	–
424	Gemma 3 27B	none	5	$0.0000	–
425	Ministral 3 3B	none	5	$0.0078	–
426	Qwen 3 0.6B	none	5	–	–
427	LFM2 1.2B	none	5	–	–
428	Gemma 3 4B	none	5	–	–
429	LFM2.5-VL-1.6B	none	5	–	–
430	Granite 4.0 350M	none	5	–	–
431	Llama 3.2 1b	none	5	–	–
432	Tiny Aya Global	none	5	–	–
433	Llama 3 8b	none	5	–	–
434	Apertus 8B Instruct	none	5	–	–
435	Granite 4.0 H 350M	none	5	–	–
436	Gemma 3n E4B Instruct	none	5	–	–
437	Gemma 3n E2B Instruct	none	5	–	–
438	Gemma 3 1B	none	5	–	–
439	NVIDIA Nemotron 3 Nano 4B	–	5	–	–
440	Gemma 3 12B	none	4	$0.0000	–
441	K2 Horizon 0.9B	–	3	–	–