$0.01 per 1M tokens.
There is no free API; the cheapest paid price is $0.01 per 1M input tokens. The weights are open, so you can run it yourself.
#75 of 200 in Embeddings.
On MTEB English v2 · Pair Classification · English, Multilingual E5 Large Instruct scores 86.2 vs 86.2 — about the same.
Through the intfloat API.
Released 2024-02-08. It is the current version in its line.
Yes, the weights are open (MIT licence).
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Multilingual E5 Large Instruct is an open-weights embeddings & retrieval model from intfloat, released Feb 2024.
EmbeddingGemma 300M (cheaper), E5 Mistral 7B Instruct (better), QZhou Embedding (open weights).
CheaperEmbeddingGemma 300MunstatedOpen+1%-80%$0.002Languages: not verifiediBetterE5 Mistral 7B InstructunstatedOpen+3%+124%$0.022Languages: loses Esperanto, es, Estonian +89+1KOpenQZhou Embeddingunstated+7%——Languages: loses Esperanto, es, Estonian +88+1
CheaperBge Base En v1.5unstatedOpensame-50%$0.005Languages: loses Esperanto, es, Estonian +89iCheaperE5 Base v2unstatedOpensame-50%$0.005Languages: loses Esperanto, es, Estonian +89
NV Embed v2unstatedOpen+3%——Languages: loses Esperanto, es, Estonian +89+1aLGAI Embedding PreviewunstatedOpen+3%——Languages: loses Esperanto, es, Estonian +89+1