Self-Hosted AI Models

153 open-source AI models you can run on your own infrastructure. Self-hosting gives you complete data privacy, zero per-token costs, and full control over the model and its behavior.

153

Open Source

Providers

125

+ Tool Use

+ Vision

+ Reasoning

Top Open Source Providers

Alibaba

32 models

DeepSeek

12 models

Zhipu AI

11 models

NVIDIA

11 models

Mistral AI

10 models

Google

9 models

Moonshot AI

8 models

All Self-Hostable Models - Ranked by Score

#	Model	Provider	Score	Context
1	DeepSeek V4 ProDeepSeek	DeepSeek	87	1.0M
2	DeepSeek V3.2DeepSeek	DeepSeek	81	164K
3	Gemma 4 31BGoogle	Google	81	262K
4	Gemma 4 31B (free)Google	Google	81	262K
5	Qwen3.5 397B A17BAlibaba	Alibaba	79	262K
6	R1 0528DeepSeek	DeepSeek	79	164K
7	GLM 5.2Zhipu AI	Zhipu AI	79	1.0M
8	GLM 5.2 (batch)Zhipu AI	Zhipu AI	78	512K
9	GLM 5.1Zhipu AI	Zhipu AI	78	205K
10	MiniMax M2.7MiniMax	MiniMax	78	205K
11	MiniMax M2.5MiniMax	MiniMax	78	205K
12	GLM 5Zhipu AI	Zhipu AI	78	205K
13	Qwen3.5-122B-A10BAlibaba	Alibaba	78	262K
14	Gemma 2 27BGoogle	Google	77	8K
15	Qwen3.5-27BAlibaba	Alibaba	77	262K
16	MiMo-V2.5-ProXiaomi	Xiaomi	76	1.1M
17	Qwen3.5-35B-A3BAlibaba	Alibaba	76	262K
18	Kimi K2.6Moonshot AI	Moonshot AI	76	262K
19	GLM 4.7Zhipu AI	Zhipu AI	75	205K
20	GLM 4.6Zhipu AI	Zhipu AI	75	205K
21	GLM 4.5Zhipu AI	Zhipu AI	75	131K
22	MiniMax M3MiniMax	MiniMax	74	1.0M
23	R1DeepSeek	DeepSeek	74	164K
24	MiniMax M3 (batch)MiniMax	MiniMax	74	524K
25	Inklingthinkingmachines	thinkingmachines	74	1.0M
26	Hy3Tencent	Tencent	74	262K
27	MiMo-V2.5Xiaomi	Xiaomi	73	1.1M
28	Gemma 4 26B A4B Google	Google	73	262K
29	Gemma 4 26B A4B (free)Google	Google	73	262K
30	Inkling (batch)thinkingmachines	thinkingmachines	73	524K
31	MiniMax M2.1MiniMax	MiniMax	72	205K
32	MiniMax M2MiniMax	MiniMax	72	205K
33	DeepSeek V3.2 ExpDeepSeek	DeepSeek	72	164K
34	DeepSeek V3.1DeepSeek	DeepSeek	72	164K
35	DeepSeek V3 0324DeepSeek	DeepSeek	72	164K
36	Inkling Smallthinkingmachines	thinkingmachines	72	524K
37	GLM 4.5 AirZhipu AI	Zhipu AI	71	131K
38	DeepSeek V3DeepSeek	DeepSeek	70	164K
39	Qwen3 VL 235B A22B ThinkingAlibaba	Alibaba	69	131K
40	Qwen3 VL 235B A22B InstructAlibaba	Alibaba	69	262K

Self-Hosting AI Models

Complete Data Privacy

Your data never leaves your infrastructure. Critical for healthcare, finance, legal, and government use cases where data residency and privacy regulations apply.

Zero Per-Token Cost

After the initial hardware investment, there are no per-request charges. At high volumes, self-hosting can be 10-100x cheaper than API-based services.

Popular Hosting Options

Run models with vLLM, Ollama, text-generation-inference, or llama.cpp. Most can run on consumer GPUs (RTX 4090) for smaller models, or cloud GPUs (A100, H100) for larger ones.

Fine-Tuning & Customization

Self-hosted models can be fine-tuned on your own data, creating domain-specific versions that outperform general-purpose models for your use case.

Frequently Asked Questions

Self-hosting gives you complete data privacy (no data leaves your servers), eliminates per-token API costs, removes rate limits, enables offline operation, and allows fine-tuning for your specific use case.

Requirements depend on model size. Small models (7B parameters) run on consumer GPUs with 8GB VRAM. Medium models (13-30B) need 24GB+ VRAM. Large models (70B+) require multiple high-end GPUs or specialized inference hardware.

Popular tools include Ollama (easiest setup), llama.cpp (most efficient for CPU inference), vLLM (fastest for GPU serving), and text-generation-webui (feature-rich UI). Each excels at different use cases.

For high-volume usage (thousands of requests/day), self-hosting is significantly cheaper. For low-volume or sporadic use, API access is more cost-effective since you avoid hardware costs and maintenance overhead.