Molecular blood biomarkers are lacking for high-grade (HG) gastroenteropancreatic (GEP) neuroendocrine neoplasms (NEN). To histologically distinguish between neuroendocrine carcinoma (NEC), neuroendocrine tumors G3 (NET G3), adenocarcinoma and MINEN is often challenging. The mRNA-based NETest has diagnostic, prognostic and predictive value in neuroendocrine tumors G1-2 but has not been studied in HG GEP-NEN. Patients with advanced HG GEP-NEN were prospectively included in an observational study. A blood sample was collected before the start of chemotherapy and pseudonymised before NETest was performed. NETest results are expressed as an activity index (NETest score) from 0 to 100. The normal score cut-off is 20. Histological sections were pseudonymised before centralized pathological re-evaluation. Samples from 60 patients were evaluable with the NETest. Main primary tumor sites were colon (14), rectum (12), pancreas (11) and esophagus (7). Re-classification: 30 NEC, 12 NET G3, 3 HG-NEN ambiguous morphology, 8 MiNEN, 3 adenocarcinomas with neuroendocrine differentiation (ADNE), 3 adenocarcinomas and 1 NET G2. Elevated NETest (>20) was seen in 38/45 (84%) HG GEP-NEN, all 17 large-cell NEC (100%), 11/13 (85%) small-cell NEC, all ambiguous cases and 7/12 (64%) NET G3. NETest was elevated in 5/8 (63%) MiNEN, 2/3 ADNE, however not in 3 adenocarcinomas. Median survival was 10.2 months (9.6–10.8 95%CI) for evaluable HG GEP-NEN treated with palliative chemotherapy (n = 39), and survival was significantly shorter in patients with NETest >60 with an OS of only 6.5 months. This is the first study to evaluate use of the NETest in advanced HG GEP-NEN. The NETest was almost always elevated in GEP-NEC and in all large-cell NEC. The NETest was also frequently elevated in NET G3 and MiNEN, however cases were limited. Baseline NETest was not predictive for benefit of chemotherapy, however a NETest >60 was prognostic with a shorter survival for patients receiving chemotherapy.