Objective To report a single‐centre experience of the regimen GAMEC (granulocyte colony‐stimulating factor, actinomycin‐D, methotrexate with folinic acid rescue, etoposide and cisplatin) over 18 years in both untreated disease and relapse settings. Methods This retrospective cohort study was based on 162 patients who received GAMEC dose‐dense chemotherapy incorporating actinomycin and high dose methotrexate. Survival outcomes were compared. Risk categorization based on (1) the International Prognostic Factor Study Group (IPFSG) criteria and (2) two factors, lactate dehydrogenase (LDH) levels greater than the upper limit of normal and age ≥35 years, were also compared in terms of survival outcomes using Cox proportional hazard regression modelling. Results Seventy‐five patients with poor‐prognosis disease, according to International Germ Cell Cancer Collaborative Group classification, received GAMEC as initial therapy. With a median follow‐up of 63 months, the median progression‐free survival (PFS) was >14 months. The 2‐year PFS rate was 61.5% (95% confidence interval [CI] 49.1–71.6), and the 3‐year overall survival (OS) rate was 71.9%. Seventy‐six patients received GAMEC as second‐line therapy (following failure of bleomycin, etoposide and cisplatin or etoposide cisplatin). The median PFS was 7.5 months (95% CI 5.2–not evaluable), the 2‐year PFS rate was 43.5% (95% CI 32.1–54.4) and the 3‐year OS rate was 53.7% (95% CI 41.6–64.3). In the third‐line setting ( n = 11), the 2‐year PFS was 18.2% (95% CI 2.8–44.2). Overall, the treatment‐related death rate declined from 10.5% in the first 15 years to 2.6% in the last 5 years. Conclusion GAMEC was an effective regimen in untreated poor‐prognosis disease and on relapse following conventional cisplatin and etoposide‐based chemotherapy. Risk categorization based on LDH/age is more sensitive than that based on the updated IPFSG criteria. It is possible to identify patients who are particularly likely to benefit from this treatment, which has the important advantages of short duration and absence of bleomycin, particularly in patients with central nervous system and mediastinal disease. Low‐dose induction treatment is associated with safer delivery of treatment without compromising survival.