Alternatives to bge-m3

bge-m3 alternatives

bge-m3's smallest published quantization is 0.59 GiB. The models below do the same job with a different trade — less memory, comparable size, or a more permissive licence.

From the file· sizes from summed file bytes

Meaningfully smaller

under 70% of its smallest quantization
ModelParamsSmallestQuantsLicence
jina-embeddings-v5-text-small596M0.19 GiB56cc-by-nc-4.0
embeddinggemma-300m-qat-q8_0-unquantized303M0.31 GiB1gemma
nomic-embed-text-v1.5137M0.05 GiB39apache-2.0
jina-embeddings-v5-text-nano212M0.09 GiB42cc-by-nc-4.0
all-MiniLM-L6-v223M0.02 GiB28apache-2.0
mxbai-embed-xsmall-v124M0.03 GiB4apache-2.0
nomic-embed-text-v2-moeMoE475M0.25 GiB20apache-2.0
snowflake-arctic-embed-l-v2.0568M0.39 GiB15

Comparable in size

within ±40%, so a like-for-like swap
ModelParamsSmallestQuantsLicence
KaLM-embedding-multilingual-mini-instruct-v2.5494M0.49 GiB1apache-2.0
Qwen3.5-9B-DFlash1.3B0.45 GiB11

More permissively licensed

licences that allow commercial use
ModelParamsSmallestQuantsLicence
KaLM-embedding-multilingual-mini-instruct-v2.5494M0.49 GiB1apache-2.0
nomic-embed-text-v1.5137M0.05 GiB39apache-2.0
all-MiniLM-L6-v223M0.02 GiB28apache-2.0
mxbai-embed-xsmall-v124M0.03 GiB4apache-2.0
nomic-embed-text-v2-moeMoE475M0.25 GiB20apache-2.0
jina-reranker-v1-tiny-en33M0.03 GiB12apache-2.0
Qwen3-Embedding-0.6B596M0.28 GiB22apache-2.0
gte-small33M0.02 GiB12mit