There is no per-unit API price: self-host only.
There is no free API. The weights are open, so you can run it yourself.
#164 of 200 in Embeddings.
On MTEB Persian v2 · Reranking · Persian, Static Similarity Mrl Multilingual v1 scores 53.3 vs 53.3 — about the same.
Released 2025-01-15. It is the current version in its line.
Yes, the weights are open (Apache-2.0 licence).
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Static Similarity Mrl Multilingual v1 is an open-weights embeddings & retrieval model from sentence-transformers, released Jan 2025.
EmbeddingGemma 300M (better), Bge M3 (open weights).
BetterEmbeddingGemma 300MunstatedOpen+34%—$0.002Languages: not verified+1
OpenBge M3unstated+40%—$0.01Languages: not verified+1iBetterMultilingual E5 LargeunstatedOpen+40%—$0.01Memory (4-bit): 0.1 GB → 0.3 GBiBetterMultilingual E5 Large InstructunstatedOpen+36%—$0.01Memory (4-bit): 0.1 GB → 0.3 GB
BetterText Embedding 3 Smallunstated+9%—$0.02Languages: not verified
Snowflake Arctic Embed L v2.0unstatedOpen+39%——Memory (4-bit): 0.1 GB → 0.3 GBcF2LLM v2 8BunstatedOpen+37%——Memory (4-bit): 0.1 GB → 4.5 GBOSolon Embeddings Large 0.1unstatedOpen+36%——Languages: loses en, es, Estonian +45+1