The Open Model Ecosystem

Llama, Mistral, Qwen, DeepSeek, Gemma and Phi are not interchangeable, and the leaderboard that ranks them is mostly measuring itself. This course teaches what open weights actually gets you, how to read the licence and the model card, how to size a model to the machine you really have, and how to settle a model choice with forty of your own examples in one afternoon.

1. 1Open Weights Is Not Open Source7 minOpen weights means you can run and adapt the model, not that you can see how it was made.
2. 2Reading the Licence Before You Ship8 minA licence travels with every fine-tune and every download; check the chain, not the badge on the page.
3. 3Hugging Face as Infrastructure8 minNever load a pickle you did not create; prefer safetensors, and pin the commit rather than the branch.
4. 4The Families, and What Each Is Actually For9 minFamilies differ durably in licence, size ladder and ecosystem; whichever leads on quality changes every few months.
5. 5Base, Instruct, Reasoning8 minReasoning models buy accuracy on multi-step problems with tokens and latency; most production tasks are not multi-step.
6. 6What Benchmarks Actually Measure9 minA benchmark result is evidence about that benchmark; transfer to your task is a hypothesis you have to test.
7. 7Choosing a Model for the Job9 minFit the constraints first, memory and latency and licence and language, then test the two or three models left standing.
8. 8Evaluate a Model on Your Task in an Afternoon10 minForty real examples and a fixed harness settle a model choice better than any leaderboard, in about four hours.