About RunItLocal

RunItLocal helps you check whether an open-weight AI model will fit in your GPU's or Mac's memory, which quantization to pick, and roughly how fast it will run.

Who runs it

RunItLocal is operated by Tilorah Core LLC. Questions or corrections: hello@tilorah.co.

Where the data comes from

Model sizes come from the publishers' model cards. Hardware memory and bandwidth come from vendor spec sheets. Speeds are estimates from our methodology, calibrated against public benchmarks:

• llama.cpp CUDA scoreboard (Llama 2 7B Q4_0, tg128, no FA)
• llama.cpp Apple Silicon benchmarks (Llama 2 7B Q4_0, tg128)

Data last updated 2026-10-15. Real-world speed depends on your software, drivers, context length and settings, so treat estimates as a guide, not a promise.

Affiliate disclosure

Some links to cloud GPU providers (Vast.ai and RunPod) are affiliate links. If you sign up through them we may earn a commission at no extra cost to you. Those links never change whether a model is shown as fitting or how fast it is estimated to run.