Open-weight LLMs you can run yourself

Updated · public data

What people are pulling from Hugging Face right now. Official releases only.

RankModelLicenseParametersMemory to runDownloadsHosted from
1Qwen3.8 27B AlibabaApache 2.028B~17 GB7.4Mhosted from $0.24
2Llama 3.1 8B MetaLlama 3.18B~5 GB6.6Mhosted from $0.05
3DeepSeek R1 DeepSeekMIT684B~414 GB1.3Mhosted from $0.7
4Gemma 4 31B GoogleApache 2.033B~20 GB9.6Mhosted from $0.2
5gpt-oss-20b OpenAIApache 2.021B~13 GB6Mhosted from $0.018
6Kimi K3 MoonshotCustom license2780B~1682 GB1.1Mhosted from $0.304
7Qwen3 0.6B AlibabaApache 2.00.8B<1 GB30.6Mnot hosted
8gpt-oss-120b OpenAIApache 2.0117B~71 GB3.4Mhosted from $0.03
9Qwen3.8 Flash Next AlibabaCustom license180B~109 GB2.3Mhosted from $0.15
10DeepSeek V4 Flash 0731 DeepSeekMIT304B~184 GB4.4Mhosted from $0.006
11GLM 5.3 Flash Z.aiMIT321B~194 GB6.7Mhosted from $0.05
12Llama 3.2 1B MetaLlama 3.21.2B<1 GB6.8Mhosted from $0.027
13Qwen2.5 7B AlibabaApache 2.07.6B~5 GB8Mhosted from $0.1
14Qwen3 8B AlibabaApache 2.08.2B~5 GB10.3Mnot hosted
15Qwen3.5 9B AlibabaApache 2.09.7B~6 GB9Mhosted from $0.1
16Meta Llama 3 8B MetaLlama 38B~5 GB793.5Knot hosted
17Qwen3.6 35B A3B AlibabaApache 2.036B~22 GB3.7Mhosted from $0.15
18Gemma 4 26B A4B GoogleApache 2.027B~16 GB11.2Mhosted from $0.09
19DeepSeek V3 DeepSeekCustom license685B~414 GB1.5Mhosted from $0.257
20DeepSeek V4.1 Flash DeepSeekMIT763B~462 GB1.2Mhosted from $0.112
21Llama 2 7B MetaLlama 26.7B~4 GB986.8Knot hosted
22Mistral 7B Instruct v0.2 Mistral AIApache 2.07.2B~4 GB1.6Mnot hosted
23Qwen2.5 VL 7B AlibabaApache 2.08.3B~5 GB5.1Mnot hosted
24GLM 5.2 Z.aiMIT753B~456 GB492.9Khosted from $0.037
25DeepSeek V3 0324 DeepSeekMIT685B~414 GB1.2Mhosted from $0.29
26Qwen3.6 27B AlibabaApache 2.028B~17 GB2.3Mhosted from $0.25
27DeepSeek V4 Pro DeepSeekMIT1599B~967 GB383.7Khosted from $0.955
28DeepSeek V3.2 DeepSeekMIT685B~415 GB4.9Mhosted from $0.259
29Llama 3.2 3B MetaLlama 3.23.2B~2 GB1.6Mhosted from $0.05
30Qwen3 VL 8B AlibabaApache 2.08.8B~5 GB7.1Mhosted from $0.25
31Qwen3.5 4B AlibabaApache 2.04.7B~3 GB8.7Mnot hosted
32GLM 5.3 Z.aiCustom license753B~456 GB1.8Mhosted from $0.45
33Qwen3 4B AlibabaApache 2.04B~2 GB12.8Mnot hosted
34Phi 2 MicrosoftMIT2.8B~2 GB530.6Knot hosted
35Mistral 7B v0.1 Mistral AIApache 2.07.2B~4 GB353.7Knot hosted
36Qwen2.5 1.5B AlibabaApache 2.01.5B<1 GB6.7Mnot hosted
37GLM 4.7 Flash Z.aiMIT31B~19 GB1.3Mhosted from $0.061
38DeepSeek V4 Flash DeepSeekMIT291B~176 GB806Khosted from $0.091
39MiniCPM5 2B OpenBMBApache 2.02.5B~2 GB1.3Mnot hosted
40Qwen3.5 35B A3B AlibabaApache 2.036B~22 GB1.6Mhosted from $0.15
41Kimi K2 MoonshotCustom license1026B~621 GB422.4Khosted from $0.57
42GLM 5 Z.aiMIT754B~456 GB815.9Khosted from $1
43Qwen2.5 0.5B AlibabaApache 2.00.5B<1 GB7.4Mnot hosted
44DeepSeek R1 Distill Qwen 1.5B DeepSeekMIT1.8B~1 GB1.3Mnot hosted
45Qwen2.5 Coder 32B AlibabaApache 2.033B~20 GB742.5Khosted from $0.66
46Qwen3 4B Instruct 2507 AlibabaApache 2.04B~2 GB3.2Mnot hosted
47Gemma 3 1B GoogleGemma1B<1 GB2Mnot hosted
48Gemma 3 4B GoogleGemma4.3B~3 GB1.1Mhosted from $0.05
49Kimi K2.5 MoonshotCustom license1027B~621 GB337Khosted from $0.5
50Llama 3.3 70B MetaLlama 3.371B~43 GB266Khosted from $0.59
51Qwen3.5 27B AlibabaApache 2.028B~17 GB1.8Mhosted from $0.26
52Qwen3.5 0.8B AlibabaApache 2.00.9B<1 GB3.4Mnot hosted
53Phi 4 MicrosoftMIT15B~9 GB363.3Khosted from $0.07
54Qwen3 14B AlibabaApache 2.015B~9 GB7.4Mhosted from $0.12
55Qwen3 32B AlibabaApache 2.033B~20 GB3.2Mhosted from $0.08
56Gemma 3 27B GoogleGemma27B~17 GB413.3Khosted from $0.08
57Qwen3 Coder Next AlibabaApache 2.080B~48 GB590.2Khosted from $0.12
58GLM 5.1 Z.aiMIT754B~456 GB499.2Khosted from $0.966
59MiniMax M2.5 MiniMaxCustom license229B~138 GB733.8Khosted from $0.27
60Qwen2.5 3B AlibabaNon-commercial3.1B~2 GB4.6Mnot hosted
61Qwen3 1.7B AlibabaApache 2.02B~1 GB4.7Mnot hosted
62Qwen3.5 397B A17B AlibabaApache 2.0403B~244 GB562.2Khosted from $0.5
63Gemma 2 2B GoogleGemma2.6B~2 GB546.6Knot hosted
64Qwen2.5 Coder 7B AlibabaApache 2.07.6B~5 GB1.9Mnot hosted
65Florence 2 Large MicrosoftMIT0.8B<1 GB360.3Knot hosted
66Qwen3 30B A3B AlibabaApache 2.031B~18 GB1.3Mhosted from $0.12
67Diffusiongemma 26B A4B GoogleApache 2.026B~16 GB669.6Knot hosted
68Qwen2 VL 7B AlibabaApache 2.08.3B~5 GB667.3Knot hosted
69Gemma 2 9B GoogleGemma9.2B~6 GB843.2Knot hosted
70DeepSeek R1 0528 DeepSeekMIT685B~414 GB172.5Khosted from $0.5
71Qwen2.5 VL 3B AlibabaCustom license3.8B~2 GB2Mnot hosted
72Kimi K2.6 MoonshotCustom license1027B~621 GB397Khosted from $0.65
73Muse Glimmer 30B MetaApache 2.030B~18 GB252.9Khosted from $0.3
74MiniMax M2.7 MiniMaxCustom license229B~138 GB657.1Khosted from $0.3
75DeepSeek R1 Distill Qwen 32B DeepSeekMIT33B~20 GB375.5Knot hosted
76SmolLM3 3B Hugging FaceApache 2.03.1B~2 GB859.5Knot hosted
77Qwen3 VL 4B AlibabaApache 2.04.4B~3 GB3.3Mnot hosted
78NVIDIA Nemotron 3 Nano 30B A3B NVIDIACustom license32B~19 GB1.2Mhosted from $0.06
79Phi 3 Mini 4k MicrosoftMIT3.8B~2 GB370.8Knot hosted
80Qwen3 30B A3B Instruct 2507 AlibabaApache 2.031B~18 GB1.1Mhosted from $0.1
81Hy3 TencentApache 2.0299B~181 GB756.3Khosted from $0.132
82Qwen3 Coder 30B A3B AlibabaApache 2.031B~18 GB422.3Khosted from $0.07
83DeepSeek R1 0528 Qwen3 8B DeepSeekMIT8.2B~5 GB525.8Knot hosted
84Qwen3.5 2B AlibabaApache 2.02.3B~1 GB3.1Mnot hosted
85SmolLM2 135M Hugging FaceApache 2.00.1B<1 GB3Mnot hosted
86Qwen3 VL 2B AlibabaApache 2.02.1B~1 GB2.4Mnot hosted
87Phi 3.5 Mini MicrosoftMIT3.8B~2 GB367Knot hosted
88Qwen3 235B A22B AlibabaApache 2.0235B~142 GB453.5Knot hosted
89Phi 3 Mini 128k MicrosoftMIT3.8B~2 GB192.5Knot hosted
90Florence 2 MicrosoftMIT0.2B<1 GB2.9Mnot hosted
91Meta Llama 3 70B MetaLlama 371B~43 GB210.1Knot hosted
92MiniMax M3 MiniMaxCustom license427B~258 GB196.8Khosted from $0.3
93MiniCPM5 1B OpenBMBApache 2.01.1B<1 GB352.2Knot hosted
94Granite Docling 258M IBMApache 2.00.3B<1 GB288.3Knot hosted
95MiniMax M2 MiniMaxCustom license229B~138 GB204.8Khosted from $0.255
96Mistral 7B Instruct v0.1 Mistral AIApache 2.07.2B~4 GB133.2Knot hosted
97DeepSeek V4 Flash Vision Exp DeepSeekMIT305B~184 GB508.5Khosted from $0.216
98MiniCPM V 4.6 OpenBMBApache 2.01.3B<1 GB279.4Knot hosted
99QwQ 32B AlibabaApache 2.033B~20 GB48.5Knot hosted
100Gemma 7B GoogleGemma8.5B~5 GB33.5Knot hosted
101DeepSeek V3.1 DeepSeekMIT685B~414 GB373.7Khosted from $0.25
102Phi 3.5 Vision MicrosoftMIT4.2B~3 GB657.8Knot hosted
103GLM 4.7 Z.aiMIT358B~217 GB85.8Khosted from $0.6
104Laguna S 2.1 PoolsideOpenMDW118B~71 GB328.1Khosted from $0.09
105DeepSeek Coder V2 Lite DeepSeekCustom license16B~10 GB705.4Knot hosted
106Qwen2 VL 2B AlibabaApache 2.02.2B~1 GB1.2Mnot hosted
107Llama 4 Scout 17B 16E MetaLlama 4109B~66 GB166.9Khosted from $0.18
108DeepSeek R1 Distill Qwen 7B DeepSeekMIT7.6B~5 GB359.7Knot hosted
109GLM 4.5 Air Z.aiMIT110B~67 GB701.9Khosted from $0.13
110Gemma 3 12B GoogleGemma12B~7 GB384.5Khosted from $0.05
111Qwen2.5 72B AlibabaCustom license73B~44 GB282Khosted from $0.36
112Llama 2 13B MetaLlama 213B~8 GB221.6Knot hosted
113Qwen3 Next 80B A3B AlibabaApache 2.081B~49 GB244.1Khosted from $0.09
114Qwen3.5 122B A10B AlibabaApache 2.0125B~76 GB677.8Khosted from $0.26
115Phi 4 Mini MicrosoftMIT3.8B~2 GB358.6Knot hosted
116GLM 4.5 Z.aiMIT358B~217 GB124.7Khosted from $0.6
117Qwen3 4B Thinking 2507 AlibabaApache 2.04B~2 GB601.3Knot hosted
118Llama 3.2 11B Vision MetaLlama 3.211B~6 GB82Knot hosted
119Kimi K2.7 Code MoonshotCustom license1027B~621 GB113.8Khosted from $0.671
120Qwen2.5 32B AlibabaApache 2.033B~20 GB1.6Mnot hosted
121Qwen2 7B AlibabaApache 2.07.6B~5 GB453.3Knot hosted
122Gemma 2B GoogleGemma2.5B~2 GB134.8Knot hosted
123NVIDIA Nemotron 3 Super 120B A12B NVIDIACustom license124B~75 GB1.1Mnot hosted
124Qwen2.5 VL 32B AlibabaApache 2.033B~20 GB776.4Knot hosted
125Step 3.5 Flash StepFunApache 2.0199B~121 GB263.2Khosted from $0.1
126DeepSeek R1 Distill Qwen 14B DeepSeekMIT15B~9 GB331Knot hosted
127Gemma 3 270m GoogleGemma0.3B<1 GB130.6Knot hosted
128Qwen2.5 14B AlibabaApache 2.015B~9 GB1.2Mnot hosted
129SmolLM2 1.7B Hugging FaceApache 2.01.7B~1 GB272.6Knot hosted
130Llama 3.1 70B MetaLlama 3.171B~43 GB128.7Khosted from $0.4
131NVIDIA Nemotron 3 Ultra 550B A55B NVIDIAOpenMDW561B~339 GB930.9Khosted from $0.6
132DeepSeek R1 Distill Llama 8B DeepSeekMIT8B~5 GB146.5Knot hosted
133Kimi Linear 48B A3B MoonshotMIT49B~30 GB321.5Knot hosted
134DeepSeek V4 Pro 0813 DeepSeekMIT1651B~999 GB141.9Khosted from $1.32
135Llama 2 70B MetaLlama 269B~42 GB22.9Knot hosted
136Qwen3 VL 30B A3B AlibabaApache 2.031B~19 GB291.2Khosted from $0.15
137Paligemma 3B Pt 224 GoogleGemma2.9B~2 GB266.8Knot hosted
138Qwen3 235B A22B Instruct 2507 AlibabaApache 2.0235B~142 GB160.6Khosted from $0.09
139Phi 1 5 MicrosoftMIT1.4B<1 GB55.7Knot hosted
140MiniCPM V 4 5 OpenBMBApache 2.08.7B~5 GB83.7Knot hosted
141Llama 3.1 Nemotron 70B NVIDIALlama 3.171B~43 GB23.9Knot hosted
142UI TARS 1.5 7B ByteDanceApache 2.08.3B~5 GB271Khosted from $0.1
143GLM 4.1V 9B Z.aiMIT10B~6 GB141.8Knot hosted
144c4ai Command R v01 CohereNon-commercial35B~21 GB67.3Knot hosted
145LFM2.5 2.6B Liquid AICustom license2.7B~2 GB114.9Knot hosted
146NVIDIA Nemotron Nano 9B v2 NVIDIACustom license8.9B~5 GB286.9Knot hosted
147Qwen 72B AlibabaCustom license72B~44 GB586.9Knot hosted
148Qwen 7B AlibabaCustom license7.7B~5 GB119.4Knot hosted
149Qwen3 Coder 480B A35B AlibabaApache 2.0480B~290 GB39.5Khosted from $0.3
150Qwen2.5 Coder 14B AlibabaApache 2.015B~9 GB1.8Mnot hosted
151MiMo V2.6 Pro XiaomiMIT1024B~620 GB99.5Khosted from $0.435
152Deepseek Coder 6.7B DeepSeekCustom license6.7B~4 GB159Knot hosted
153DeepSeek V2 DeepSeekCustom license236B~143 GB41.3Knot hosted
154MiniCPM V 4.7 35B A3B OpenBMBApache 2.035B~21 GB9.3Knot hosted
155d1 3B Liquid AICustom license3.1B~2 GB16.7Knot hosted
156Deepseek Coder 1.3B DeepSeekCustom license1.4B<1 GB58.8Knot hosted
157Deepseek Coder 33B DeepSeekCustom license33B~20 GB4.7Knot hosted
158LFM2.5 24B A2B Liquid AICustom license24B~15 GB6.5Khosted from $0.05
159MiniCPM V 4.7 1B OpenBMBApache 2.01.6B<1 GB5.3Knot hosted
160Deepseek Math 7B DeepSeekCustom license6.9B~4 GB11.3Knot hosted
161d1 Omni 600M Liquid AICustom license0.6B<1 GB12.9Knot hosted
162Muse Spark Open 816B A42B MetaCustom license822B~497 GB402hosted from $1.25
163NV Reason CT NVIDIAOpenMDW5.3B~3 GB5.5Knot hosted
164Ling 3.1 Flash Ant GroupMIT572B~346 GB666not hosted
165AesCode 8B MicrosoftApache 2.08.8B~5 GB476not hosted
166AesCode 32B MicrosoftApache 2.033B~20 GB503not hosted
167NVIDIA Nemotron 3.5 Bolt 74B A9B 24GB NVIDIAOpenMDW22B~14 GB58not hosted
168NVIDIA Nemotron 3.5 Super VL 120B A12B NVIDIAOpenMDW124B~75 GB146not hosted
169Flux 3 Image Klein Pup Black Forest LabsApache 2.09.4B~6 GB13not hosted
170NVIDIA Nemotron 3.5 Bolt 74B A9B NVIDIAOpenMDW75B~45 GB4not hosted

Open-weight image models you can run yourself · Open-weight video models you can run yourself

Data: Hugging Face (official accounts) · hosted prices: OpenRouter · memory is a rough estimate · Oct 16, 2026