Granite Embedding 30M
A small IBM embedding model for semantic search. The 30M variant is English-only; the 278M variant is multilingual. Both take up to 512 tokens.
Exact local Ollama manifest and configuration checked on 2026-10-20. Memory figures are conservative planning estimates; no DeviceTerra inference or speed test is claimed. 512 token maximum. 30M English only; 278M multilingual. Generates vectors, not chat answers. Embeddings return vectors, not chat answers.
- Official model source
- Official engine artifact
Will it run?
MamiLens uses 9 GB as a conservative planning floor for this profile and its listed quantization. This is not measured speed. Processor, backend support, context length, memory bandwidth and other applications still affect the result.
Check your complete system →What is it useful for?
Suggested Ollama artifact
Confirm the tag on the official page before downloading, then run:
ollama pull granite-embedding:30mStart with a representative non-sensitive task. Record the exact runtime, quantization, context, load time, generation speed, memory use and answer quality.
Evidence and limits
Published context: 512 token input limit. The source establishes specifications; it does not guarantee that maximum context fits at the MamiLens minimum RAM floor.