The reasoning_effort values OpenAI accepts, in ascending order.
Confirmed against the live API rather than taken from docs: the provider's rejection
message for an unsupported value enumerates this exact set. xhigh and none sit outside MJ's
numeric 1-100 effort scale by design — see OpenAILLM.getReasoningLevel.
The
reasoning_effortvalues OpenAI accepts, in ascending order.Confirmed against the live API rather than taken from docs: the provider's rejection message for an unsupported value enumerates this exact set.
xhighandnonesit outside MJ's numeric 1-100 effort scale by design — see OpenAILLM.getReasoningLevel.