OpenAI Alternatives: Comparing Retrieval Models for RAG - Cohere Embed v3 and Open Source

Explore a comprehensive comparison of top retrieval models for Retrieval Augmented Generation (RAG) in this 17-minute video. Dive into the performance of OpenAI's text-embedding-ada-002, Cohere's new Embed v3, and the open-source e5-base-v2 model. Examine MTEB leaderboards, inference speeds, and query results across different models. Analyze the strengths and weaknesses of each option, with a particular focus on the differences between Ada 002 and Cohere v3. Gain valuable insights into selecting the most suitable embedding model for your RAG applications through practical tests and expert analysis.

MTEB Leaderboards
Starting with OpenAI, Cohere, and e5
Inference Speeds
Querying with Different Models
Results between models
Ada 002 vs Cohere v3
Another test for OpenAI, Cohere, and E5
More Questions and Final Thoughts