LOCAL OR HOSTED

Start with where you want the model to run.

HostingLinks separates downloadable open-weight models from hosted assistants so hardware estimates, privacy trade-offs and model comparisons stay honest.

RUN IT YOURSELF

Local AI

Ollama, LM Studio, downloadable weights, VRAM, RAM, privacy and hardware planning.

Explore local AI

PROVIDER-HOSTED

Claude, GPT and Gemini

Current hosted model families, what each tier is for, and where you can access them.

Browse hosted AI

CHOOSE THE APPROACH

Local vs hosted

Compare control, convenience, hardware, privacy and ongoing cost.

Compare the options