diff --git a/service/openaicompat/chat_to_responses.go b/service/openaicompat/chat_to_responses.go index 76aa6d25d..c6cdc6944 100644 --- a/service/openaicompat/chat_to_responses.go +++ b/service/openaicompat/chat_to_responses.go @@ -323,6 +323,10 @@ func ChatCompletionsRequestToResponsesRequest(req *dto.GeneralOpenAIRequest) (*d if req.MaxCompletionTokens > maxOutputTokens { maxOutputTokens = req.MaxCompletionTokens } + // OpenAI Responses API rejects max_output_tokens < 16 when explicitly provided. + if maxOutputTokens > 0 && maxOutputTokens < 16 { + maxOutputTokens = 16 + } var topP *float64 if req.TopP != 0 {