Local model requirements
What hardware do I need to run this local AI model?
Pick a model size, quantization level, and machine type to estimate the RAM, VRAM, and storage needed before buying hardware or downloading a giant model file.
Buyer rule
Buy for the model class, not the brand name.
A 70B Q4 model and an 8B Q4 model live in completely different hardware worlds.
Model file
~6GB
Keep extra disk room for multiple downloads and updates.
Minimum memory
~8GB
Includes estimated runtime and context overhead.
Comfortable memory
~14GB
Leaves room for the OS, browser, and local AI app.
Use this before buying a machine for local AI
The estimate is deliberately plain-English. Model architecture, context length, quantization quality, and runner support can change the exact number, but these thresholds are useful for avoiding underpowered purchases.
AI cost desk
AI Model Pricing Sheet
A worksheet for comparing AI provider costs, hidden pricing drivers, model fit, and budget assumptions without relying on stale static prices.
Provider cost worksheet plus budget notes. Updated when major pricing changes ship.
Use the calculator