gemma-3-4b-it — VINDEX3

google/gemma-3-4b-it encoded as a VINDEX3 container for LARQL: the original BF16 weights, unmodified, plus the model's structure as an executable, inspectable graph.

larql pull larql/gemma-3-4b-it
larql run gemma-3-4b-it "What is the capital of France?"
larql serve gemma-3-4b-it --port 8080      # OpenAI-compatible API

Supported: text generation (run, chat, serve, LQL, vindex3 observe). The vision weights are carried in the container, but this LARQL build does not admit image or audio input for this model.

Encoded from google/gemma-3-4b-it@093f9f388b31de276ce2de164bdc2081324b9767. Verification and digests are in PROVENANCE.md.

License

Gemma is provided under and subject to the Gemma Terms of Use found at https://ai.google.dev/gemma/terms. Use of this model must comply with the Gemma Prohibited Use Policy at https://ai.google.dev/gemma/prohibited_use_policy.