clevis@tinyweights:/home$ cat ./welcome.txt
> Small language models like Gemma, Phi, SmolLM, and Qwen, run locally. Benchmarks, quantization, and hands-on deployment guides for small LLMs on real hardware.
> New releases tested the week they drop. Straight talk on what is worth running and what is not.
46 posts · last updated 2026-07-19 · all writing CC BY 4.0
clevis@tinyweights:/home$ ls -lh --sort=time | grep guides
07-14
8min
guides
Run Bonsai 27B Locally: A 27B Model in 3.9 GB, Even on a Phone
07-08
8min
guides
How to Run North Mini Code Locally: Cohere's 30B-A3B Coding Model
07-07
9min
guides
Local LLM for Private Document Q&A: Build an Offline RAG Pipeline
07-02
7min
guides
How to Quantize a Model with llama.cpp: From Safetensors to GGUF
07-01
9min
guides
DiffusionGemma: Google's Text-Diffusion Model That Generates 4x Faster, and How to Run It Locally
06-30
8min
guides
How to Run LFM2.5-8B-A1B Locally: Liquid AI's On-Device Tool-Calling MoE
06-08
6min
guides
How to Run Phi-4-mini-reasoning Locally: Microsoft's 3.8B Math Model
06-07
7min
guides
How to Run Qwen3.6-35B-A3B Locally: A 35B MoE Model on One GPU
06-04
6min
guides
Run IBM Granite 4.1 Locally: 3B, 8B, and 30B Setup Guide
06-01
12min
guides
1-bit LLMs Explained: How BitNet's Ternary Weights Actually Work