Try Delta
← All models

Meta2024

Llama 3.2

Dense · vision on 11B class

Meta’s on-device Llama: 1B and 3B for edge chat, with the family known for vision-enabled siblings. Light enough for laptops, familiar instruction style.

  • Fits fromPhone / edge · 4 GB+
  • Contextup to 131K
  • Sizes2
  • RunsLocal on Delta
VisionEdgeMultilingualOn-device LlamaLightweight chatEdge kits

Sizes

Llama 3.2 1B

1.3 GB • 131k ctx • Q8_0

Llama 3.2 3B

2.3 GB • 131k ctx • Q5_K_M

Memory at context length

Approximate working set for Llama 3.2 1B as context grows.

4K
1.6 GB
8K
1.7 GB
16K
1.8 GB
33K
2 GB
66K
2.1 GB
131K
2.2 GB

Related families

Llama 3.1

Meta’s 8B generalist for chat and writing: the classic Llama stack, fully local.

Ministral 3

Mistral’s edge family with native vision: phones through workstations, fully local.

Gemma 3N

Google’s efficient open model, tuned to run locally on phones, tablets, and laptops.