Text, image, video, speech and embedding models behind OpenAI- and Anthropic-compatible APIs, routed by task difficulty.
检索与聚类，批量接口自动合并请求。
model/embed-1024