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.
#72 of 89 in Multimodal.
On MIEB lite · Document Understanding · English, Siglip Base Patch16 256 Multilingual scores 22.9 vs 25 — lower.
Released 2024-01-08. It is the current version in its line.
Yes, the weights are open (Apache-2.0 licence).
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Siglip Base Patch16 256 Multilingual is an open-weights embeddings & retrieval model from Google, released Jan 2024.
Jina Embeddings v5 Omni Nano (better), Argus Colqwen3.5 9B v0 (open weights).
BetterJina Embeddings v5 Omni NanounstatedOpen×3.2—$0.02Memory (4-bit): 0.2 GB → 0.6 GBDOpenArgus Colqwen3.5 9B v0unstated×3.9——Memory (4-bit): 0.2 GB → 5.3 GB
Nemotron Colembed VL 8B v2Open×3.9——Memory (4-bit): 0.2 GB → 5.2 GBDArgus Colqwen3.5 4B v0unstatedOpen×3.9——Memory (4-bit): 0.2 GB → 2.8 GBvVultronRetrieverPrime Qwen3.5 8BOpen×3.9——Memory (4-bit): 0.2 GB → 5 GBvVultronRetrieverCore Qwen3.5 4.5BOpen×3.9——Memory (4-bit): 0.2 GB → 2.7 GB