What is a cheaper alternative to OpenAI embeddings?
retained text
We are on OpenAI's embedding API today. It was fine until we started re-indexing nightly, at which point the per-token cost stopped being a rounding error. We are looking at TIE and Infinity. But open to other options as well. May vLLM too. Need advise.
document fetched 2026-09-22 00:29 · canonical https://reddit.com/r/selfhosted/comments/1wm2fmp/what_is_a_cheaper_alternative_to_openai_embeddings