Available endpoints and models

The following four endpoints are available:

• /v1/models
• /v1/chat/completions
• /v1/embeddings
• /rerank

For the chat completions endpoint, the model gpt-oss-120b is currently available. This is a pure text model.

The embeddings endpoint uses the model bge-m3, which converts texts into vector representations.

The rerank endpoint uses the model bge-reranker-v2-m3. This model is employed together with bge-m3 and its purpose is to select the most relevant top‑K results from the hits generated by the embedding model.