A compact, multilingual embedding specialist that punches above its weight for its size. It converts text into dense vector representations across multiple languages, handling long contexts up to 32k tokens — useful when you need to embed lengthy documents without chunking. Being a 'small' variant, it trades some raw capability for speed and lower resource usage.