Background: The biological behavior of primary small gastric gastrointestinal stromal tumor (gGIST) is indolent. The cutoff size categorizing small gGIST continues to be controversial. To date, there is no consensus regarding whether it should be 1 cm, 2 cm, or another size. We aimed to find a new cutoff size.Methods: Retrospective clinicopathological and prognosis data of patients with small gGIST from January 1998 to January 2015 were collected among five medical centers in southern China. Tumor size was divided into two groups: <1 cm (Mirco group) and 1-2 cm (Small group). We compared the clinicopathological index and prognosis between these two groups and identified a new cutoff size to define small gGIST.Results: During this 18-year period, there were 276 patients with primary small gGIST treated at these five medical centers. The range of tumor size was 0.2-2.0 cm. The median tumor size was 1.0 cm. The range of the mitotic count was 0-70/50 high power fields (HPFs) with counts ≤5/50 HPFs in 259 patients (93.8%), 5< counts ≤10/50 HPFs in 7 patients (2.5%), and counts >10/50 HPFs in 10 patients (3.6%). The median follow-up time was 38 months (3-156 months). The 5-year overall survival rate was 98.7% in the entire group. Using Pearson correlation analysis, there was a positive correlation between the mitotic count and tumor size as a continuous variable (r=0.164, P=0.006). There were 137 patients in the Micro group and 139 cases in the Small group. In the Micro group, mitotic counts were ≤5/50 HPFs in 134 patients, 5< counts ≤10/50 HPFs in 0 patients, and counts >10/50 HPFs in 3 patients; mitotic counts in the Small group were counts ≤5/50 HPFs in 125 patients, 5< counts ≤10/50 HPFs in 7 patients, >10/50 HPFs in 7 patients.There was a statistically significant difference between these two groups (P=0.002); the Small group had more intermediate/high-risk cases. Using the receiver operating characteristic curve (ROC curve), we observed that 1.15 cm was the new cutoff size to separate low-risk cases and intermediate/high-risk cases (AUC =0.707, P=0.004, sensitivity =0.824, 1-specificity =0.429).Conclusions: Primary small gGIST has a good prognosis; gGIST <1 cm can be regarded as benign tumors that only requires endoscopic ultrasonography (EUS) follow-up. The proportion of potential intermediate/ high-risk disease is high for patients with 1-2 cm gGIST. These patients should be treated with caution and the tumors should be resected if necessary. These results indicate that 1.15 cm may be the new cutoff size to separate small gGIST from large gGIST, but further studies are needed for verification.