Choose an embedding approach from your retrieval needs.
—
Plugsky is OpenAI-compatible with flat-rate plans — see pricing or start on the free plan (2 free models, no card).
The Embedding Model Comparison page lays out how embedding models differ by vector dimensions, maximum input length, cost and benchmark scores, so you can choose one for semantic search, clustering or retrieval-augmented generation. It suits developers and data teams picking an embedding backend, where dimension count drives index size and recall, and maximum input decides how much text fits in a single vector. Compare candidates, then test retrieval quality on your own documents before committing.
Higher dimensions can capture more nuance but increase index size, memory and search cost. Many teams start with a compact model and move up only if retrieval quality on their own data falls short.
Yes. Vectors from different models are not compatible, even at the same dimension count. Switching models means re-embedding the whole corpus and rebuilding the index.
No. Benchmarks are a useful filter, but retrieval quality depends on your domain, language and query style. Test the shortlist on a sample of your own documents and real queries.
Canonical pricing and plans: plugsky.com/#sec-pricing · Terms · SLA · Docs