6 min readLocal & Private AI
How To Choose Between Two Local Models Without Guessing
A bigger model is not automatically the better one for rewriting. A fifteen minute test using your own writing that settles it properly.
Read article →Topic
Open Weights
33 articles tagged Open Weights. For the wider topic, see Local & Private AI.
6 min readLocal & Private AI
How To Keep A Local Model Updated Without Breaking Your Workflow
Local models do not auto-update, which is a feature. When a new version is worth the download, how to test it against your own writing, and when to skip it.
Read article →6 min readLocal & Private AI
What Open Weights Does Not Mean
Open weights is not open source, not a promise of privacy, and not a licence to do anything. What the term actually covers, and the four things people wrongly assume.
Read article →6 min readLocal & Private AI
Who Is Still Shipping Open Weights in August 2026
The open model supply has concentrated sharply this year. A dated scoreboard of who is publishing, under which licenses, and what it means for procurement.
Read article →6 min readLocal & Private AI
llama.cpp Passed 100,000 Stars: The Infrastructure Under Local AI
The project that made local models practical crossed a milestone in 2026. What it actually does, why it matters for privacy, and what it means that it is a dependency.
Read article →6 min readLocal & Private AI
Mistral's New Open-Weight Family: Why a European Publisher Still Matters
Mistral put a new open-weight model into early access without disclosing specifications. Why the European supply question matters separately from benchmarks.
Read article →6 min readLocal & Private AI
DeepSeek V4 Flash 0731: MIT Licensed, Cheap, and Not for Your Laptop
DeepSeek shipped an MIT-licensed 284B mixture-of-experts model on 31 July 2026. What the license actually gives you, and why cheap hosting is the real story.
Read article →6 min readLocal & Private AI
GGUF Explained for People Who Just Want to Run a Model
What GGUF files are, how to read the cryptic names, what Q4_K_M actually means, and the three things to check before you download several gigabytes.
Read article →7 min readLocal & Private AI
How to Benchmark a Local Model on Your Own Writing
Public benchmarks measure coding and mathematics. Here is a twenty-minute test that measures whether a model will actually help with your email.
Read article →6 min readLocal & Private AI
How to Run an Open-Weights Model on Windows in 2026
Three paths from nothing to a working local model, what each one costs you, and the five settings that determine whether it feels fast or unusable.
Read article →6 min readLocal & Private AI
Int4 Quantization in 2026: How Much Quality Do You Actually Lose?
Four-bit quantization is now the default for local models. What it costs in quality, where the floor is, and why the answer depends entirely on your task.
Read article →6 min readLocal & Private AI
Mixture of Experts, Explained Without the Maths
Nearly every large model in 2026 is a mixture of experts. What that means, why the two parameter counts matter differently, and the planning mistake it causes.
Read article →6 min readLocal & Private AI
Open-Weights Models for Non-English Professional Writing
Small models handle some languages far better than others. How to test coverage on your own text, and which families are worth trying first.
Read article →6 min readLocal & Private AI
The Best Open-Weights Models for Writing Work in 2026
A practical shortlist of open-weights models that actually help with professional writing, sorted by what hardware you have rather than by benchmark score.
Read article →6 min readLocal & Private AI
Can You Run a Trillion-Parameter Model Locally? An Honest Answer
Open weights at frontier scale sound like they put frontier AI on your desk. The arithmetic says otherwise. What the memory math actually allows, and what you should run instead.
Read article →6 min readLocal & Private AI
DeepSeek V4: What an Efficiency-First Open Model Changes
DeepSeek V4 competes on inference efficiency rather than raw benchmark scores. Why that is the more interesting strategy, and what it means for anyone paying for AI by the token.
Read article →6 min readLocal & Private AI
Gemma 4 for Local Writing: Google's On-Device Bet
Gemma 4 moved to Apache 2.0 and ships sizes built for laptops rather than datacenters. What it is good at, where it fits against Qwen, and why on-device is the interesting category.
Read article →6 min readLocal & Private AI
GLM 5.2 and Why Agentic Reliability Is a Different Skill
Zhipu's GLM line built its reputation on function calling and structured output rather than benchmark scores. Why that reliability is harder to achieve than raw capability, and where it matters.
Read article →6 min readLocal & Private AI
gpt-oss: What OpenAI's Open-Weights Models Are Actually For
OpenAI ships gpt-oss under Apache 2.0 in 120B and 20B sizes. Where they fit, why a reasoning-oriented model is a mixed blessing for rewriting, and how they compare to the small Qwen and Gemma models.
Read article →6 min readLocal & Private AI
How to Choose an Open-Weights Model for Your Laptop
A decision procedure that starts with your actual RAM instead of a leaderboard. Which size, which variant, which quantization, and how to know when you have picked wrong.
Read article →6 min readLocal & Private AI
How to Read an AI Model Card (and a System Card)
Model cards and system cards are the closest thing to a datasheet AI has. What to look for, what the omissions tell you, and why this is becoming a compliance document.
Read article →6 min readLocal & Private AI
Inkling and the Return of US Open-Weights Releases
Thinking Machines Lab shipped its first open-weights model in July 2026. Why a US entry matters in a field that had tilted heavily toward Chinese labs.
Read article →6 min readLocal & Private AI
Kimi K3 and the Trillion-Parameter Open-Weights Era
Moonshot AI released Kimi K3 as a 2.8-trillion-parameter open-weights model in July 2026. What that actually means, who can run it, and why it changes the field even if you never touch it.
Read article →6 min readLocal & Private AI
MiniMax M3, Sparse Attention, and the Cost of Long Context
MiniMax M3 handles a million-token context at a fraction of standard transformer compute. How sparse attention works in plain terms, and why long context is cheaper but not better.
Read article →6 min readLocal & Private AI
Mistral, Open Models, and European Data Sovereignty
For European organizations, where a model runs is a compliance question. How Mistral's open lineup fits, and why sovereignty is solved by architecture more often than by geography.
Read article →6 min readLocal & Private AI
NVIDIA Nemotron and the Art of Compressing a Big Model
NVIDIA keeps shipping compressed open-weight models derived from larger ones. How pruning and distillation work, and why compressed models are the ones that reach your laptop.
Read article →6 min readLocal & Private AI
What Open-Weights Benchmarks Do Not Tell You About Writing
SWE-bench, MMLU-Pro, and AIME scores predict almost nothing about whether a model will rewrite your email well. What to measure instead.
Read article →6 min readLocal & Private AI
Open-Weights Model Licenses Explained: Apache, MIT, and Community Terms
Apache 2.0, MIT, Llama community licenses, and research-only terms give you very different rights. A plain-English guide to what you can legally do with a downloaded model.
Read article →7 min readLocal & Private AI
Open Weights vs Cloud APIs: The 2026 Cost Comparison Nobody Runs
Cloud inference prices fell roughly 80 percent between 2025 and 2026. Here is what that does to the local-versus-cloud calculation, and why cost is now the wrong reason to choose either one.
Read article →6 min readLocal & Private AI
Open Weights vs Open Source AI: What the Labels Actually Mean
Open weights, open source, and open license are three different claims. What each one gives you, what it withholds, and which one matters if you want AI that runs on your own machine.
Read article →6 min readLocal & Private AI
The Qwen Open-Weights Family, Explained for People Who Just Want It to Work
Alibaba ships Qwen models faster than anyone can track. A guide to which sizes matter, what the naming means, and why the small ones are the important ones.
Read article →6 min readLocal & Private AI
Tencent's Hunyuan 3.0 and the Rise of Selectable Inference Modes
Hunyuan 3.0 shipped open weights in July 2026 with three selectable inference modes. Why letting the user choose how hard the model thinks is the most practical feature of 2026.
Read article →7 min readLocal & Private AI
Why Open Weights Matter for Workplace Privacy
Open weights turn a data-handling promise into an architectural fact. Why that distinction is the whole argument for anyone writing client, patient, or contract text.
Read article →