Comment by cahaya

Wondering how this compares to the Gemini (preview) embeddings as they seem to perform significantly better than OpenAI embeddings 3 large. I don't see any MTEB scores so hard to compare.