Classifica

Classifica: Open weights

Modelli open weights per Podium Score.

Generale700 Coding399 Matematica53 Reasoning413 Agentico55 Conoscenza412 Multimodale382 Agenti di codice55 Reasoning scientifico82 Multilingue12 Scrittura61 Chat e assistenti61 Ricerca85 Chiamata di strumenti32 Contesto lungo82 Velocità402 Valore401 Open weights170

170 modelli

#	Modello	Podium Score	Arena Elo	Intelligenza	Programmazione	Velocità	Prezzo (uscita)
01	Kimi K3 Moonshot AI · Aperto · reasoning	83.3	1485	57.0	61.5	37 t/s	$15/M
02	Qwen3.8 Flash Next Alibaba · Aperto	66.8	1460	61.5	60.5	168 t/s	$0.4/M
03	G9v3-39A5B AI9Stars · Aperto	65.0	—	30.0	—	60 t/s	$2/M
4	G9v3-3B AI9Stars · Aperto	65.0	—	30.0	—	60 t/s	$2/M
5	HyperNova 60B Multiverse · Aperto	65.0	—	30.0	—	367 t/s	$0.14/M
6	LFM2.5-2.6B Liquid AI · Aperto	65.0	—	30.0	—	60 t/s	$2/M
7	LFM2.5-8B-A1B Liquid AI · Aperto	65.0	—	30.0	—	338 t/s	$2/M
8	LFM2.5-VL-1.6B Liquid AI · Aperto	65.0	—	30.0	—	395 t/s	$2/M
9	LongCat 2.0 LongCat · Aperto	65.0	—	30.0	—	42 t/s	$2.95/M
10	Magistral Small 1.2 Mistral · Aperto	65.0	—	30.0	—	60 t/s	$1.5/M
11	Qwen3.8 27B (xhigh) Alibaba · Aperto	65.0	—	30.0	—	53 t/s	$3/M
12	Qwen3.8 27B (low) Alibaba · Aperto	65.0	—	30.0	—	64 t/s	$3/M
13	Qwen3.8 27B (medium) Alibaba · Aperto	65.0	—	30.0	—	62 t/s	$3/M
14	Qwen3.8 27B (Non-reasoning) Alibaba · Aperto · reasoning	65.0	—	30.0	—	62 t/s	$3/M
15	GLM-5.3 Flash Z.ai · Aperto	64.2	1445	58.0	58.4	182 t/s	$0.3/M
16	GLM-5.2 Z.ai · Aperto	59.1	1469	51.0	41.9	164 t/s	$4.29/M
17	Granite 4.2 30B Instruct IBM · Aperto	58.6	1380	51.0	56.0	78 t/s	$1.2/M
18	Qwen3 8 2 4t A95b Alibaba · Aperto	57.7	—	—	40.0	85 t/s	$2/M
19	DeepSeek V4.1 Flash DeepSeek · Aperto · reasoning	54.5	1450	40.0	52.0	214 t/s	$1.2/M
20	Glm 5 Z.ai · Aperto	52.6	—	—	40.0	60 t/s	$3/M
21	Sarvam 105B Sarvam AI · Aperto	50.0	—	—	—	60 t/s	$3/M
22	Sarvam 30B Sarvam AI · Aperto	50.0	—	—	—	60 t/s	$3/M
23	Granite 4.2 8B Instruct IBM · Aperto	49.2	1315	43.0	44.0	115 t/s	$0.45/M
24	Kimi K2.6 Moonshot AI · Aperto	49.0	1461	—	36.4	40 t/s	$4/M
25	Sarvam M Reasoning Sarvam AI · Aperto · reasoning	49.0	—	—	—	40 t/s	$3/M
26	MiMo V2.5 Pro Xiaomi · Aperto	48.0	1466	42.0	42.6	65 t/s	$0.87/M
27	DeepSeek V4 Pro DeepSeek · Aperto	46.0	1458	44.0	30.7	66 t/s	$0.87/M
28	GLM-5.1 Z.ai · Aperto	46.0	—	—	57.0	74 t/s	$4.4/M
29	Deepseek V4 Pro 0424 DeepSeek · Aperto	45.3	—	—	40.0	60 t/s	$1.1/M
30	Deepseek V4 Pro 0424 High DeepSeek · Aperto	43.7	—	—	40.0	60 t/s	$1.1/M
31	Deepseek V4 Flash 0420 DeepSeek · Aperto	42.1	—	—	25.0	60 t/s	$1.1/M
32	DeepSeek V4 Flash DeepSeek · Aperto	42.0	1438	50.0	38.0	113 t/s	$0.28/M
33	Granite 4.2 3B Instruct IBM · Aperto	42.0	1260	36.0	35.0	148 t/s	$0.2/M
34	Kimi K2.7 Code Moonshot AI · Aperto	41.8	—	42.0	39.9	39 t/s	$4/M
35	Qwen3 6 Max Alibaba · Aperto	41.1	—	—	40.0	85 t/s	$2/M
36	Deepseek V4 Flash 0420 High DeepSeek · Aperto	39.0	—	—	25.0	60 t/s	$1.1/M
37	Deepseek V3 2 Reasoning DeepSeek · Aperto · reasoning	38.0	—	—	40.0	40 t/s	$1.1/M
38	Mimo V2 5 0424 Xiaomi · Aperto	38.0	—	—	40.0	60 t/s	$3/M
39	Qwen3 6 27b Alibaba · Aperto	37.7	—	—	40.0	85 t/s	$2/M
40	Glm 5 1 Non Reasoning Z.ai · Aperto · reasoning	36.3	—	—	40.0	40 t/s	$3/M
41	Gemma 4 31B Google · Aperto	35.0	1451	29.0	—	35 t/s	$0.3/M
42	Glm 5 2 Non Reasoning Z.ai · Aperto · reasoning	34.8	—	—	40.0	40 t/s	$3/M
43	Glm 5 Non Reasoning Z.ai · Aperto · reasoning	33.2	—	—	40.0	40 t/s	$3/M
44	Qwen3 5 397b A17b Non Reasoning Alibaba · Aperto · reasoning	32.7	—	—	40.0	55 t/s	$2/M
45	Minimax M2 1 MiniMax · Aperto	32.1	—	—	25.0	60 t/s	$2/M
46	Qwen3 6 35b A3b Alibaba · Aperto	32.1	—	—	40.0	85 t/s	$2/M
47	Deepseek V4 Pro 0424 Non Reasoning DeepSeek · Aperto · reasoning	31.9	—	—	40.0	40 t/s	$1.1/M
48	Mimo V2 Flash Reasoning Xiaomi · Aperto · reasoning	31.9	—	—	25.0	40 t/s	$3/M
49	Qwen3 5 Omni Plus Alibaba · Aperto	31.3	—	—	40.0	85 t/s	$2/M
50	Qwen3 6 27b Non Reasoning Alibaba · Aperto · reasoning	31.3	—	—	40.0	55 t/s	$2/M
51	Deepseek V3 1 Terminus Reasoning DeepSeek · Aperto · reasoning	31.1	—	—	40.0	40 t/s	$1.1/M
52	GPT-oss-120B OpenAI · Aperto	30.0	—	—	37.0	240 t/s	$0.4/M
53	Qwen3 5 27b Non Reasoning Alibaba · Aperto · reasoning	30.0	—	—	40.0	55 t/s	$2/M
54	Deepseek V4 Flash Non Reasoning DeepSeek · Aperto · reasoning	29.3	—	—	25.0	40 t/s	$1.1/M
55	Glm 4 6 Reasoning Z.ai · Aperto · reasoning	29.3	—	—	40.0	40 t/s	$3/M
56	Minimax M2 MiniMax · Aperto	28.9	—	—	25.0	60 t/s	$2/M
57	Mimo V2 5 Pro Non Reasoning Xiaomi · Aperto · reasoning	28.4	—	—	40.0	40 t/s	$3/M
58	Qwen3 5 122b A10b Non Reasoning Alibaba · Aperto · reasoning	28.2	—	—	40.0	55 t/s	$2/M
59	Glm 4 7 Non Reasoning Z.ai · Aperto · reasoning	27.1	—	—	40.0	40 t/s	$3/M
60	Gemma 4 26b A4b Google · Aperto	26.1	—	—	40.0	120 t/s	$5/M
61	Deepseek V3 2 Reasoning 0925 DeepSeek · Aperto · reasoning	25.9	—	—	40.0	40 t/s	$1.1/M
62	Qwen3 Max Thinking Preview Alibaba · Aperto · reasoning	25.5	—	—	40.0	55 t/s	$2/M
63	Qwen3 6 35b A3b Non Reasoning Alibaba · Aperto · reasoning	24.6	—	—	40.0	55 t/s	$2/M
64	Qwen3 5 35b A3b Non Reasoning Alibaba · Aperto · reasoning	24.3	—	—	40.0	55 t/s	$2/M
65	Qwen3.5 397B-A17B Alibaba · Aperto	23.9	1442	34.0	11.5	71 t/s	$2/M
66	Gemma 4 31b Non Reasoning Google · Aperto · reasoning	22.3	—	—	40.0	80 t/s	$5/M
67	Gemma 4 12b Google · Aperto	22.2	—	—	40.0	120 t/s	$5/M
68	Qwen3.5 122B-A10B Alibaba · Aperto	22.0	1417	32.0	—	126 t/s	$2/M
69	GPT-oss-20B OpenAI · Aperto	22.0	—	—	27.0	90 t/s	$0.2/M
70	Qwen3 5 9b Alibaba · Aperto	21.8	—	—	40.0	85 t/s	$2/M
71	Deepseek V3 2 0925 DeepSeek · Aperto	21.7	—	—	40.0	60 t/s	$1.1/M
72	Qwen3 Coder Next Alibaba · Aperto	21.3	—	—	40.0	85 t/s	$2/M
73	Deepseek V3 1 Reasoning DeepSeek · Aperto · reasoning	21.0	—	—	40.0	40 t/s	$1.1/M
74	Qwen3 Vl 235b A22b Reasoning Alibaba · Aperto · reasoning	20.9	—	—	40.0	55 t/s	$2/M
75	Qwen3 5 9b Non Reasoning Alibaba · Aperto · reasoning	20.6	—	—	40.0	55 t/s	$2/M
76	Gemma 4 26b A4b Non Reasoning Google · Aperto · reasoning	20.4	—	—	40.0	80 t/s	$5/M
77	Qwen3 5 4b Alibaba · Aperto	20.4	—	—	40.0	85 t/s	$2/M
78	Qwen3 235b A22b Instruct 2507 Reasoning Alibaba · Aperto · reasoning	19.9	—	—	40.0	55 t/s	$2/M
79	Mistral Small 4 Mistral · Aperto	19.7	—	—	40.0	100 t/s	$6/M
80	Qwen3 Max Alibaba · Aperto	19.6	1435	—	—	85 t/s	$2/M
81	Qwen3 Max Preview Alibaba · Aperto	19.4	—	—	40.0	85 t/s	$2/M
82	Qwen3 5 Omni Flash Alibaba · Aperto	19.2	—	—	25.0	85 t/s	$2/M
83	Nemotron 3 Ultra 550B NVIDIA · Aperto	19.0	1426	38.0	19.1	142 t/s	$2.75/M
84	Deepseek R1 0120 DeepSeek · Aperto · reasoning	18.6	—	—	40.0	40 t/s	$1.1/M
85	Qwen3 235b A22b Instruct 2507 Alibaba · Aperto	18.4	—	—	40.0	85 t/s	$2/M
86	Qwen3 Coder 480b A35b Instruct Alibaba · Aperto	18.2	—	—	40.0	85 t/s	$2/M
87	Qwen3 Vl 32b Reasoning Alibaba · Aperto · reasoning	18.1	—	—	40.0	55 t/s	$2/M
88	Minimax M1 80k MiniMax · Aperto	17.9	—	—	25.0	60 t/s	$2/M
89	Glm 4 6v Reasoning Z.ai · Aperto · reasoning	16.9	—	—	40.0	40 t/s	$3/M
90	Qwen3 Next 80b A3b Reasoning Alibaba · Aperto · reasoning	16.9	—	—	40.0	55 t/s	$2/M
91	Qwen3 5 4b Non Reasoning Alibaba · Aperto · reasoning	16.1	—	—	40.0	55 t/s	$2/M
92	Mistral Large 3 Mistral · Aperto	15.9	—	—	40.0	100 t/s	$6/M
93	Glm 4 7 Flash Non Reasoning Z.ai · Aperto · reasoning	15.6	—	—	25.0	40 t/s	$3/M
94	Deepseek V3 0324 DeepSeek · Aperto	15.2	—	—	40.0	60 t/s	$1.1/M
95	Mistral Small 3 1 Mistral · Aperto	14.9	—	—	40.0	100 t/s	$6/M
96	Mistral Medium 3 1 Mistral · Aperto	14.7	—	—	40.0	100 t/s	$6/M
97	Qwen3 30b A3b 2507 Reasoning Alibaba · Aperto · reasoning	14.6	—	—	40.0	55 t/s	$2/M
98	Qwen3 Vl 235b A22b Instruct Alibaba · Aperto	14.4	—	—	40.0	85 t/s	$2/M
99	Qwen3 Next 80b A3b Instruct Alibaba · Aperto	13.8	—	—	40.0	85 t/s	$2/M
100	Qwen3 Coder 30b A3b Instruct Alibaba · Aperto	13.6	—	—	40.0	85 t/s	$2/M
101	Qwen3 235b A22b Instruct Reasoning Alibaba · Aperto · reasoning	13.5	—	—	40.0	55 t/s	$2/M
102	Qwen3 Vl 30b A3b Reasoning Alibaba · Aperto · reasoning	13.4	—	—	40.0	55 t/s	$2/M
103	Gemma 4 12b Non Reasoning Google · Aperto · reasoning	13.2	—	—	40.0	80 t/s	$5/M
104	DeepSeek V3.2 DeepSeek · Aperto	12.9	1425	—	35.5	60 t/s	$1.1/M
105	Mistral Medium 3 Mistral · Aperto	12.5	—	—	40.0	100 t/s	$6/M
106	Llama Nemotron Super 49b V1 5 Reasoning Meta · Aperto · reasoning	12.4	—	—	40.0	50 t/s	$0/M
107	Mistral Small 4 Non Reasoning Mistral · Aperto · reasoning	12.3	—	—	40.0	60 t/s	$6/M
108	Gemma 4 E4b Google · Aperto	12.2	—	—	40.0	120 t/s	$5/M
109	Qwen3 32b Instruct Reasoning Alibaba · Aperto · reasoning	11.4	—	—	40.0	55 t/s	$2/M
110	Qwen3 Vl 32b Instruct Alibaba · Aperto	11.0	—	—	40.0	85 t/s	$2/M
111	Qwen3 235b A22b Instruct Alibaba · Aperto	10.8	—	—	40.0	85 t/s	$2/M
112	Mistral Small 3 2 Mistral · Aperto	10.7	—	—	40.0	100 t/s	$6/M
113	Qwen3 Vl 8b Reasoning Alibaba · Aperto · reasoning	10.5	—	—	25.0	55 t/s	$2/M
114	Qwen3 14b Instruct Reasoning Alibaba · Aperto · reasoning	10.4	—	—	25.0	55 t/s	$2/M
115	Llama 4 Scout Meta · Aperto	10.3	—	—	40.0	80 t/s	$0/M
116	Hermes 4 Llama 3 1 70b Reasoning Meta · Aperto · reasoning	9.9	—	—	40.0	50 t/s	$0/M
117	Qwen3 Vl 30b A3b Instruct Alibaba · Aperto	9.9	—	—	40.0	85 t/s	$2/M
118	Deepseek R1 Distill Llama 70b DeepSeek · Aperto · reasoning	9.8	—	—	40.0	40 t/s	$1.1/M
119	Qwen3 Omni 30b A3b Reasoning Alibaba · Aperto · reasoning	9.5	—	—	40.0	55 t/s	$2/M
120	Llama 3 3 Instruct 70b Meta · Aperto	9.3	—	—	40.0	80 t/s	$0/M
121	Qwen3 30b A3b Instruct Reasoning Alibaba · Aperto · reasoning	9.2	—	—	40.0	55 t/s	$2/M
122	Glm 4 5v Reasoning Z.ai · Aperto · reasoning	9.0	—	—	40.0	40 t/s	$3/M
123	Llama 3 1 Nemotron Ultra 253b V1 Reasoning Meta · Aperto · reasoning	8.9	—	—	40.0	50 t/s	$0/M
124	Qwen3 30b A3b 2507 Alibaba · Aperto	8.9	—	—	40.0	85 t/s	$2/M
125	Hermes 4 Llama 3 1 405b Reasoning Meta · Aperto · reasoning	8.8	—	—	40.0	50 t/s	$0/M
126	Gemma 4 E4b Non Reasoning Google · Aperto · reasoning	8.7	—	—	40.0	80 t/s	$5/M
127	Hermes 4 Llama 3 1 405b Meta · Aperto	8.6	—	—	40.0	80 t/s	$0/M
128	Llama Nemotron Super 49b V1 5 Meta · Aperto	8.5	—	—	40.0	80 t/s	$0/M
129	Qwen3 32b Instruct Alibaba · Aperto	8.5	—	—	40.0	85 t/s	$2/M
130	Qwen3 8b Instruct Reasoning Alibaba · Aperto · reasoning	8.3	—	—	25.0	55 t/s	$2/M
131	Qwen3 Vl 8b Instruct Alibaba · Aperto	8.2	—	—	25.0	85 t/s	$2/M
132	Gemma 3 27b Google · Aperto	7.4	—	—	40.0	120 t/s	$5/M
133	Llama 3 1 Instruct 8b Meta · Aperto	7.4	—	—	25.0	80 t/s	$0/M
134	Llama 3 1 Nemotron Instruct 70b Meta · Aperto	7.4	—	—	40.0	80 t/s	$0/M
135	Mistral Large 2407 Mistral · Aperto	7.0	—	—	40.0	100 t/s	$6/M
136	Qwen3 14b Instruct Alibaba · Aperto	6.8	—	—	25.0	85 t/s	$2/M
137	Hermes 4 Llama 3 1 70b Meta · Aperto	6.7	—	—	40.0	80 t/s	$0/M
138	Mistral Small 3 Mistral · Aperto	6.7	—	—	40.0	100 t/s	$6/M
139	Qwen3 30b A3b Instruct Alibaba · Aperto	6.6	—	—	40.0	85 t/s	$2/M
140	Llama 3 1 Instruct 70b Meta · Aperto	6.5	—	—	40.0	80 t/s	$0/M
141	Qwen Turbo Alibaba · Aperto	6.0	—	—	40.0	85 t/s	$2/M
142	Phi 4 Mini Microsoft · Aperto	5.7	—	—	25.0	60 t/s	$0.6/M
143	Gemma 3 12b Google · Aperto	5.5	—	—	40.0	120 t/s	$5/M
144	DeepSeek R1 DeepSeek · Aperto · reasoning	5.4	1422	—	—	60 t/s	$1.1/M
145	Granite 4 0 H Small IBM · Aperto	4.9	—	—	40.0	60 t/s	$3/M
146	Hermes 3 Llama 3 1 70b Meta · Aperto	4.8	—	—	40.0	80 t/s	$0/M
147	Qwen3 8b Instruct Alibaba · Aperto	4.8	—	—	25.0	85 t/s	$2/M
148	Qwen3 Omni 30b A3b Instruct Alibaba · Aperto	4.8	—	—	40.0	85 t/s	$2/M
149	Mistral Small Mistral · Aperto	4.3	—	—	40.0	100 t/s	$6/M
150	Mistral Large Mistral · Aperto	4.1	—	—	40.0	100 t/s	$6/M
151	Llama 2 Chat 7b Meta · Aperto	3.9	—	—	25.0	80 t/s	$0/M
152	Mistral Medium Mistral · Aperto	3.2	—	—	40.0	100 t/s	$6/M
153	Mistral Small 2402 Mistral · Aperto	3.2	—	—	40.0	100 t/s	$6/M
154	Llama 3 Instruct 70b Meta · Aperto	3.1	—	—	40.0	80 t/s	$0/M
155	Llama 3 2 Instruct 11b Vision Meta · Aperto	3.0	—	—	40.0	80 t/s	$0/M
156	Olmo 3 7b Instruct AI2 · Aperto	2.4	—	—	25.0	60 t/s	$3/M
157	Mistral 7b Instruct Mistral · Aperto	1.7	—	—	25.0	100 t/s	$6/M
158	Granite 3 3 8b Instruct IBM · Aperto	1.3	—	—	25.0	60 t/s	$3/M
159	Gemma 3 1b Google · Aperto	1.0	—	—	25.0	120 t/s	$5/M
160	Gemma 3 4b Google · Aperto	1.0	—	—	40.0	120 t/s	$5/M
161	Gemma 3n E2b Google · Aperto	1.0	—	—	40.0	120 t/s	$5/M
162	Gemma 3n E4b Google · Aperto	1.0	—	—	40.0	120 t/s	$5/M
163	Llama 3 Instruct 8b Meta · Aperto	1.0	—	—	25.0	80 t/s	$0/M
164	Falcon 2 TII · Aperto	0.0	—	—	40.0	60 t/s	$3/M
165	Minimax H3 MiniMax · Aperto	0.0	—	—	25.0	60 t/s	$2/M
166	Mistral Mistral · Aperto	0.0	—	—	40.0	100 t/s	$6/M
167	Qwen Image Alibaba · Aperto	0.0	—	—	40.0	85 t/s	$2/M
168	Qwen3 Asr Alibaba · Aperto	0.0	—	—	40.0	85 t/s	$2/M
169	Apodex 1.1 Independent · Aperto	0.0	—	—	—	145 t/s	$0.8/M
170	Quasar 438B Open Source · Aperto · reasoning	0.0	—	—	—	95 t/s	$1.6/M

Il Podium Score è una combinazione ponderata di cinque classifiche. Consulta la Metodologia per i dettagli.