Models

LLMBase publishes a curated inference catalog. Use the stable model ID shown below in POST /v1/chat/completions requests.

Loading models from the API...

How to use the catalog

Choose by capability first and price second. For example, use models with tools for function calling, models with structured_outputs for schema-bound JSON, and models with cache-read pricing for repeated long prompts.

For model-list endpoints, metadata fields, filtering examples, and SDK discovery code, see Model discovery.

Only inference models are accessible via agents. This model catalog is the source for direct inference model IDs; for subscription-backed agents that use a llmbase_chat_... key, see Models for agents.

Fixed location Germany

Prices in the catalog are the normal prices. On request, we provide a dedicated endpoint with fixed location Germany. It costs the normal price plus a 20% surcharge on all token prices (input, cached input and output), i.e. 1.2x each listed price.

For example, a model listed at $1.00 / 1M input tokens costs $1.20 / 1M input tokens on the Germany endpoint. Request it via Enterprise.