| Model | Context | Input $/M | Output $/M | Cache Read $/M | Cache Write $/M |
|---|
vLLM, LM Studio, MLX, Ollama (beyond the recommended pulls below), OpenAI-Compatible, LiteLLM, Google Vertex AI, Microsoft Azure AI Foundry, and AWS Bedrock connect to whatever endpoint or deployment you point them at, so Loophole doesn't pin static pricing for those — model lists are autodetected from your endpoint.