API proxy service for MBZUAI-IFM/K2-Think model with token management and load balancing.
K2Think provides an API proxy gateway for the MBZUAI-IFM/K2-Think model. It offers token rotation and load balancing, automatic failure detection and retry mechanisms, token pool management, and supports OpenAI Function Calling tool integration.
Key endpoints include:
/v1/chat/completionsfor chat completions/v1/modelsfor model information/healthfor service health checks
Administrative endpoints allow monitoring and management of tokens, including token statistics, reset functions, consecutive failure tracking, and token updater controls.