Skip to content
Tokenly.
Demo fixture — not an offerDemo fixture

Llama 3.2 3B Instruct

meta-llama/Llama-3.2-3B-Instruct

A compact 3-billion-parameter instruction model from Meta intended for lightweight assistants where latency and cost matter more than depth.

Llama 3.2 3B Instruct
Model IDmeta-llama/Llama-3.2-3B-Instruct
DeveloperMeta
Serving providerPending confirmation
Capabilitieschat
Modalitiestext
Context window128K tokens
Pricing
Input
Pending verification
Output
Pending verification
LimitationsSmall model. Expect weaker multi-step reasoning than larger siblings.
SourceModel card ↗
Last verified18 Nov 2026