Open-Source & Local LLMs
Open-weight models and running AI locally on your own hardware.
- Mistral Large 4 Goes Open-Weight: 1 Trillion Parameters and What It Takes to Run LocallyMistral released the open weights of its 1-trillion-parameter Large 4 on October 29, 2026; the 2-bit build fits a 512 GB Apple Silicon Ultra machine, while full-speed serving uses eight H200 GPUs.
- How to Run an LLM Locally: Ollama vs llama.cpp, Hardware Needs and Quantisation ExplainedOllama and llama.cpp let you run open-weight language models on your own computer, keeping prompts private. Here is how to check your GPU, estimate memory from quantisation, install either tool and set the context window and network settings safely.