[FEAT] Embeddings Benchmark

We're currently looking for a solution similar to this, but for embeddings model.

Creating this issue to give visibility to this feature for the repo

The issue does not name any files, tests, or entry points, and only requests an embeddings-model benchmark similar to an unspecified solution. First inspect the repository's existing benchmark structure and define the supported models, measurements, and comparison output; the work is done when those requirements and validation tests are established.

Written by the indexing model from the issue text.