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.
#174 of 200 in Embeddings.
On MTEB Mandarin Chinese v1 · Classification · Mandarin Chinese, M2V Base Output scores 43.2 vs 43.3 — about the same.
Released 2024-09-21. 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.
M2V Base Output is an open-weights embeddings & retrieval model from minishlab, released Sep 2024.
Qwen 3 Embedding 8B (better), QZhou Embedding (open weights).
BetterQwen 3 Embedding 8BunstatedOpen+78%—$0.01Languages: not verified+1KOpenQZhou Embeddingunstated+85%——Memory (4-bit): 0 GB → 4.2 GB
Youtu EmbeddingOpen+82%——Memory (4-bit): 0 GB → 1.6 GBKQZhou Embedding ZhOpen+82%——Languages: not verified+1iJasper Token Compression 600MunstatedOpen+80%——Memory (4-bit): 0 GB → 0.4 GBKKaLM Embedding Multilingual Mini Instruct v2.5unstatedOpen+79%——Memory (4-bit): 0 GB → 0.3 GB