Local AI guides
Use these compact guides when you need a fast answer before opening the full calculator. Each guide is generated from the same model and hardware dataset as the main tool.
Small local chat, assistant, and starter-agent picks for mainstream 8GB GPUs.
Coding-focused local model choices for RTX 3060 12GB / RTX 4070-class setups.
Pick Ollama commands by clean-fit VRAM tier before testing in the calculator.
Decide when to run locally, when to accept offload tradeoffs, and when to compare cloud/API options.
32GB Blackwell local AI guide with clean-fit picks, Qwen3.6 27B, and larger-model caveats.
VRAM fit, Ollama command, and 24GB/32GB guidance for Qwen3.6 27B.