diff --git a/common/redis.go b/common/redis.go
index 1a05721c..12c477b8 100644
--- a/common/redis.go
+++ b/common/redis.go
@@ -61,3 +61,8 @@ func RedisDel(key string) error {
 	ctx := context.Background()
 	return RDB.Del(ctx, key).Err()
 }
+
+func RedisDecrease(key string, value int64) error {
+	ctx := context.Background()
+	return RDB.DecrBy(ctx, key, value).Err()
+}
diff --git a/controller/relay-text.go b/controller/relay-text.go
index e8dab514..761ca86f 100644
--- a/controller/relay-text.go
+++ b/controller/relay-text.go
@@ -194,7 +194,11 @@ func relayTextHelper(c *gin.Context, relayMode int) *OpenAIErrorWithStatusCode {
 	if err != nil {
 		return errorWrapper(err, "get_user_quota_failed", http.StatusInternalServerError)
 	}
-	if userQuota > 10*preConsumedQuota {
+	err = model.CacheDecreaseUserQuota(userId, preConsumedQuota)
+	if err != nil {
+		return errorWrapper(err, "decrease_user_quota_failed", http.StatusInternalServerError)
+	}
+	if userQuota > 100*preConsumedQuota {
 		// in this case, we do not pre-consume quota
 		// because the user has enough quota
 		preConsumedQuota = 0
diff --git a/model/cache.go b/model/cache.go
index 64666c86..55fbba9b 100644
--- a/model/cache.go
+++ b/model/cache.go
@@ -95,6 +95,14 @@ func CacheUpdateUserQuota(id int) error {
 	return err
 }
 
+func CacheDecreaseUserQuota(id int, quota int) error {
+	if !common.RedisEnabled {
+		return nil
+	}
+	err := common.RedisDecrease(fmt.Sprintf("user_quota:%d", id), int64(quota))
+	return err
+}
+
 func CacheIsUserEnabled(userId int) bool {
 	if !common.RedisEnabled {
 		return IsUserEnabled(userId)