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.
#176 of 200 in Embeddings.
On MTEB Polish v1 · STS · Polish, Venusaur scores 52.3 vs 52.8 — about the same.
Released 2024-04-29. 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.
Venusaur is an open-weights embeddings & retrieval model from Mihaiii, released Apr 2024.
Gte Base (better), Jina Embeddings v3 (open weights).
OpenJina Embeddings v3unstated+39%—$0.05Memory (4-bit): 0 GB → 0.3 GBiBetterE5 Base v2unstatedOpen+12%—$0.005Memory (4-bit): 0 GB → 0.1 GB
BetterBge Base En v1.5unstatedOpen+9%—$0.005Memory (4-bit): 0 GB → 0.1 GBiBetterMultilingual E5 Large InstructunstatedOpen+34%—$0.01Memory (4-bit): 0 GB → 0.3 GB
Gte Qwen2 7B InstructunstatedOpen+35%——Memory (4-bit): 0 GB → 4.2 GBsMmlw Roberta BaseunstatedOpen+35%——Memory (4-bit): 0 GB → 0.1 GB