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.
#117 of 200 in Embeddings.
On MTEB multilingual v2 · Reranking · multilingual (22 languages), Bekko Embedding v1 A8m scores 60.5 vs 60.6 — 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 A8m is an open-weights embeddings & retrieval model from hotchpotch, released Jul 2026.
EmbeddingGemma 300M (better), F2LLM v2 8B (open weights).
BetterEmbeddingGemma 300MunstatedOpen+4%—$0.002Max input: 8K tokens → 2K tokens+1cOpenF2LLM v2 8Bunstated+16%——Memory (4-bit): 0.1 GB → 4.5 GB
BetterQwen 3 Embedding 8BunstatedOpen+8%—$0.01Memory (4-bit): 0.1 GB → 4.5 GB
BetterBge M3unstatedOpen+4%—$0.01Memory (4-bit): 0.1 GB → 0.3 GBiBetterMultilingual E5 Large InstructunstatedOpen+3%—$0.01Max input: 8K tokens → 1K tokens+1
Llama Embed Nemotron 8BOpen+12%——Memory (4-bit): 0.1 GB → 4.5 GBOOcten Embedding 8BOpen+12%——Memory (4-bit): 0.1 GB → 4.5 GB