Model
Inference & Serving
Llama
1Meta's 2open 3family 4of 5language 6models.
the six Ws · specification
W1
Who
Released by Meta.
W2
What
A family of open-weight text and multimodal models across many sizes.
W3
Where
Weights run locally or via providers.
W4
When
When you want a capable open base model.
W5
Why
The most widely deployed open model family.
W6
With
A runtime such as vLLM, Ollama, or Transformers.
W7
Watch
Community license with some use restrictions; weights gated on Hugging Face, run them yourself.
alternatives