Feedback
Allow OpenAI-compatible requests to use the server’s output-token limit
- Status
- Fixed
- Fixed in
- v26.11.6
- Area
- compiler
- Source discussion
- -
- Last updated
- 2026-11-19
Upvotes
0 upvotes
Uses your Objo forum account.
Public summary
Objo apps using LLM.OpenAICompatible cannot set LLMRequestOptions.MaximumOutputTokens to 0 to use the server's configured generation limit. A zero value fails validation before a request is sent. It should leave the output-token limit to the server.