Ilblur【出乎意料】OpenClaw 上模型 TOP 3 :gpt-5.3-codex gemini-3-flash-preview minimax-m2.1 中发帖

模型排位是按 Percentage of  tasks  completed successfully across standardized OpenClaw agent tests 。 
出乎意料的是,opus-4.6 只排位第八,都不在前五,kimi-k2.5 排在 sonnet-4.5 之前也没想到。
来源: