Small Language Models Under 4GB: What Actually Works?

Never get stuck without AI again. Run three Small Language Models (SLMs)—also called Local LLMs—TinyLlama, Gemma-3 and Phi-4-mini—completely offline; all fit in 4 GB or less and work on any laptop and older hardware. ──────────────────── 🔧 Hardware & Software used • Laptop Ryzen 5 4500U, 8GB RAM, Ollama (no GPU needed!) • Phone iPhone 13 Pro with Mobile PocketPal AI (local GGUF) ──────────────────── 🔗 Model resources • ChatGPT global outage (news) https://timesofindia.indiatimes.com/etimes/trending/openais-chatgpt-down-globally-users-flooded-with-error-messages/articleshow/121752441.…