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.
#67 of 89 in Multimodal.
On MIEB lite · Image Clustering · English, Clip Vit Large Patch14 scores 71.3 vs 72.2 — about the same.
Released 2021-02-26. It is the current version in its line.
Yes, the weights are open.
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Clip Vit Large Patch14 is an open-weights embeddings & retrieval model from OpenAI, released Feb 2021.
EmbeddingGemma 2 (better), EVA02 CLIP B 16 (open weights).
BetterEmbeddingGemma 2unstatedOpen+10%—$0.02QOpenEVA02 CLIP B 16unstated+22%——Memory (4-bit): 0.3 GB → 3 GB
BetterJina Embeddings v5 Omni SmallunstatedOpen+17%—$0.05Memory (4-bit): 0.3 GB → 1 GBVBetterVoyage Multimodal 3+7%—$0.12eBetterEbind FullunstatedOpen+15%——Memory (4-bit): 0.3 GB → 1.1 GB
Dinov2 LargeOpen+15%——Max input: not verified