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.
#71 of 200 in Embeddings.
On MTEB Portuguese v1 · Multilabel Classification · Portuguese, SFR Embedding 2 R scores 22.9 vs 22.9 — about the same.
Released 2024-06-14. It is the current version in its line.
Yes, the weights are open (cc-by-nc-4.0 licence).
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
SFR Embedding 2 R is an open-weights embeddings & retrieval model from Salesforce, released Jun 2024.
Gemini Embedding 001 (better), Giga Embeddings Instruct 3B 0826 (open weights).
BetterGemini Embedding 001unstated+23%—$0.15Max input: 4K tokens → 2K tokensaOpenGiga Embeddings Instruct 3B 0826unstated+50%——
BetterHarrier OSS v1 27BOpen+67%——Memory (4-bit): 4.3 GB → 16 GBaBetterGiga Embeddings Instruct 10B A1.8B 0826unstatedOpen+49%——Memory (4-bit): 4.3 GB → 6.3 GB
BetterKaLM Embedding Gemma3 12B 2511Open+38%——Languages: not verified+1
Llama Embed Nemotron 8BOpen+27%——Languages: not verified+1