Download Latest Version 0.48.0_ Several models at once source code.zip (419.9 kB)
Home / 0.45.0
Name Modified Size InfoDownloads / Week
Parent folder
appcast.xml 2026-10-18 25.8 kB
Llama.dmg 2026-10-18 1.3 MB
0.45.0_ Embedding models source code.tar.gz 2026-10-17 302.1 kB
0.45.0_ Embedding models source code.zip 2026-10-17 404.7 kB
README.md 2026-10-17 799 Bytes
Totals: 5 Items   2.0 MB 0

Llama now runs embedding models, starting with EmbeddingGemma 2. They turn text into vectors for search and retrieval, through the server's /v1/embeddings endpoint. They carry an Embedding chip, and their model page links to a guide on how to use them.

  • Sign in with Hugging Face to download gated models
  • Say when a gated model needs access instead of blaming the network
  • Add a Load row to the model page to warm a model up before use
  • Let decision models take inputs up to 4k tokens on every Mac
  • Update the engine to b11671, adding support for Nemotron 3.5 Super VL
  • Fit Qwen 3.8 Flash Next on 128 GB Macs
  • Show the Liquid AI logo for LFM models
  • Use standard-size controls and a one-line token field in Settings
  • Show focus on the shortcut recorder and cancel it on a click outside
Source: README.md, updated 2026-10-17