According to CoinMeta, Wu said that the decentralized AI infrastructure project Mira launched a AI inference service Sterling on November 5th, which is currently in the beta testing phase. This service evaluates different AI API service providers by continuously monitoring indicators such as model identity, tool calls, parameter execution, abnormal responses, and context length, and prioritizes assigning user requests to those service providers that have passed verification. Sterling currently supports calling Claude, GPT, and Gemini through a unified API key. During the launch period, the prices of some models varied by up to 50% from the official prices, and it also supports pay-as-you-go usage with USDC or Mira tokens via X402.