Enterprise agent platform for building, deploying, and operating AI teammates.
Xinference chat priority
When a running model advertises supports_chat_priority: true, Xagent
automatically includes priority_hint: -1 in streaming and non-streaming
Xinference chat requests made by interactive commands. Xinference uses lower
values for earlier scheduling. An explicit request priority takes precedence.
Batch commands and automatic resumes use normal priority by default. Automatic resumes can be classified as interactive for Xagent's own task queue while still using normal priority in Xinference. Models without the advertised capability receive no automatic priority hint.