Embedding Models
Every production embedding and reranker model with pricing, dimensions, max input tokens, MTEB score, and licensing. The catalog every RAG agent needs and nobody else publishes in machine-readable form.
Embeddings are the bottleneck nobody talks about. The model is whatever; the embedding is forever. If you switch from OpenAI 3-small to Voyage 3, you re-embed your entire corpus. So the choice deserves more than a five-minute scan, and the data deserves to be machine-readable.
Sort:
For agents: same data at /api/embeddings. Filter by ?type=embedding or ?type=reranker. Free, no auth, cached 10 min.