Hugging Face models you can deploy as an API
ModelDock launches a Hugging Face model behind an OpenAI-compatible API. These are popular open models, all public and at or under the 10B-parameter limit. Paste any other public model link in the dashboard to launch it too. Every launch starts as a free 15-minute test; see the quickstart and pricing.
| MODEL | SIZE | LICENSE | CONTEXT | TYPICAL GPU | OUTPUT / 1M TOKENS |
|---|---|---|---|---|---|
| Qwen3 8B Alibaba Qwen | 8.2B | apache-2.0 | 40,960 tokens | 24 GB | $1.44 – $2.29 |
| Qwen3 4B Alibaba Qwen | 4B | apache-2.0 | 40,960 tokens | 16 GB | $1.21 |
| Qwen3 1.7B Alibaba Qwen | 2B | apache-2.0 | 40,960 tokens | 16 GB | $1.21 |
| Qwen2.5 7B Instruct Alibaba Qwen | 7.6B | apache-2.0 | 32,768 tokens | 24 GB | $1.44 – $2.29 |
| Qwen2.5 Coder 7B Instruct Alibaba Qwen | 7.6B | apache-2.0 | 32,768 tokens | 24 GB | $1.44 – $2.29 |
| DeepSeek R1 Distill 7B DeepSeek | 7.6B | mit | 131,072 tokens | 24 GB | $1.44 – $2.29 |
| Mistral 7B Instruct Mistral AI | 7.2B | apache-2.0 | 32,768 tokens | 24 GB | $1.44 – $2.29 |
| Phi-4 mini Microsoft | 3.8B | mit | 131,072 tokens | 16 GB | $1.21 |
| SmolLM3 3B Hugging Face | 3.1B | apache-2.0 | 65,536 tokens | 16 GB | $1.21 |
| Granite 3.3 8B Instruct IBM | 8.2B | apache-2.0 | 131,072 tokens | 24 GB | $1.44 – $2.29 |
| OLMo 2 7B Instruct Ai2 | 7.3B | apache-2.0 | 4,096 tokens | 24 GB | $1.44 – $2.29 |
// facts from Hugging Face as of 2026-10-25. GPU tier and price are Pilot’s typical pick for 16-bit weights, not a guarantee; paid launches are opening soon. Most downloaded: 11.3M.