llm-router

LLM Router is a drop-in OpenAI-compatible API designed to reduce LLM operational costs by automatically routing requests between different models including GLM, Kimi, Qwen, and Doubao. By using the model identifier "auto", the system selects a model based on the task type: simple tasks are routed to Doubao, balanced tasks to Qwen, and complex reasoning tasks to Kimi or GLM. Key capabilities include: - Multi-model routing across various providers. - Cost optimization to reduce API spending. - OpenAI-compatible API for easy integration without significant code changes. - Automatic fallback mechanisms to handle failures. - Task-based smart model selection.