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.
#72 of 200 in Embeddings.
On MTEB English v2 · Pair Classification · English, NoInstruct Small Embedding v0 scores 85.1 vs 85.1 — about the same.
Released 2024-05-01. 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.
NoInstruct Small Embedding v0 is an open-weights embeddings & retrieval model from avsolatorio, released May 2024.
EmbeddingGemma 300M (better), QZhou Embedding (open weights).
BetterEmbeddingGemma 300MunstatedOpen+3%—$0.002Languages: not verified+1KOpenQZhou Embeddingunstated+9%——Memory (4-bit): 0 GB → 4.2 GB
BetterBge En IclunstatedOpen+4%—$0.01Languages: not verified+1
BetterQwen 3 Embedding 8BunstatedOpen+3%—$0.01Languages: not verified+1
NV Embed v2unstatedOpen+4%——Memory (4-bit): 0 GB → 4.7 GBaLGAI Embedding PreviewunstatedOpen+4%——Memory (4-bit): 0 GB → 4.3 GB