fix(billing): settle tiered retries with final group (#6518)

relay/common/relay_info.go

@@ -122,7 +122,7 @@ type RelayInfo struct {
 	// 必须在提交前锁定全额。
 	ForcePreConsume bool
 	// Billing 是计费会话，封装了预扣费/结算/退款的统一生命周期。
-	// 免费模型时为 nil。
+	// 初始免费组可为 nil；若 auto 重试切换到付费组，会在发送前创建。
 	Billing BillingSettler
 	// BillingSource indicates whether this request is billed from wallet quota or subscription.
 	// "" or "wallet" => wallet; "subscription" => subscription
@@ -163,8 +163,9 @@ type RelayInfo struct {
 	// It is surfaced onto the consume/task log's admin_info for auditing.
 	QuotaClamp *common.QuotaClamp
 
-	// TieredBillingSnapshot is a frozen snapshot of tiered billing rules
-	// captured at pre-consume time. Non-nil only when billing mode is "tiered_expr".
+	// TieredBillingSnapshot captures tiered billing rules at pre-consume time.
+	// Auto-group retries refresh its group-dependent fields before each attempt
+	// and again before settlement. Non-nil only when billing mode is "tiered_expr".
 	TieredBillingSnapshot *billingexpr.BillingSnapshot
 	BillingRequestInput   *billingexpr.RequestInput