fix: consolidate chat-completions compatibility fixes

- apply default mapped model only when scheduling fallback is actually used

- preserve reasoning in OpenAI-compatible output via reasoning_content and avoid invalid input function_call ids

backend/internal/handler/openai_chat_completions.go

@@ -181,13 +181,7 @@ func (h *OpenAIGatewayHandler) ChatCompletions(c *gin.Context) {
 		service.SetOpsLatencyMs(c, service.OpsRoutingLatencyMsKey, time.Since(routingStart).Milliseconds())
 		forwardStart := time.Now()
 
-		defaultMappedModel := ""
-		if apiKey.Group != nil {
-			defaultMappedModel = apiKey.Group.DefaultMappedModel
-		}
-		if fallbackModel := c.GetString("openai_chat_completions_fallback_model"); fallbackModel != "" {
-			defaultMappedModel = fallbackModel
-		}
+		defaultMappedModel := c.GetString("openai_chat_completions_fallback_model")
 		result, err := h.gatewayService.ForwardAsChatCompletions(c.Request.Context(), c, account, body, promptCacheKey, defaultMappedModel)
 
 		forwardDurationMs := time.Since(forwardStart).Milliseconds()