Heavyweight 8-billion-parameter variant served through Ollama on a single GPU-capable host, fronted by an OpenWebUI chat interface. Requires roughly 16 GB of RAM and is tuned for longer-context reasoning tasks.
Ollama (Llama 3.1 8B)
Open WebUI
Run DeepSeek locally with OpenWebUI for advanced reasoning and coding assistance, fully offline.
Google's Gemma 3 (1B) language model served locally via Ollama with OpenWebUI - a compact, efficient open-weights model suitable for resource-friendly on-device AI.
Run Meta Llama 3.2 locally with OpenWebUI for a private, fully offline AI chat experience.
Run Alibaba Qwen locally with OpenWebUI for a private, multilingual AI assistant.