There is no per-unit API price: API price not public.
There is no free API. The weights are open, so you can run it yourself.
#107 of 200 in Embeddings.
On MTEB French v1 · Retrieval · multilingual (2 languages), Bekko Embedding v1 A25m scores 53.6 vs 53.8 — about the same.
Released 2026-07-28. 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.
Bekko Embedding v1 A25m is an open-weights embeddings & retrieval model from hotchpotch, released Jul 2026.
Multilingual E5 Large Instruct (better), F2LLM v2 8B (open weights).
BetterJina Embeddings v3unstatedOpen+11%—$0.05Memory (4-bit): 0.1 GB → 0.3 GBcBetterF2LLM v2 14BunstatedOpen+35%——Memory (4-bit): 0.1 GB → 8.4 GB