How to pick the right LLM before you build

TLDR DevOps shared a GitHub repo called whichllm, a curated list of open-source large language models ranked by performance. The idea is simple: don't guess which model fits your project — look at the scores.

The repo sorts models by benchmarks like MMLU, HumanEval, and coding tasks. It includes models from Meta, Mistral, Google, and others. There's a quick-start section that shows how to run them locally with tools like Ollama or LM Studio.

The author is a developer who builds side projects. The list is updated regularly. No paywall. No vendor lock-in.

This is the kind of resource the cousin who runs the side business needs. The one who tried three APIs before settling on the right one. The one who doesn't want to get burned by a model that sounds good in marketing but chokes on real work.

The repo also links to comparison pages and community discussions. It's not a product pitch. It's a map.

Why this matters for us: when you're building a tool for la comunidad — a WhatsApp bot, a small business dashboard, a legal aid helper — picking the wrong model is the most expensive mistake you'll make, and this list saves you that cost.