Neurosurgeons use three main surgical approaches for left-sided glioblastoma (GB) in eloquent areas: biopsy, tumor resection under general anesthesia (GA), and awake craniotomy (AC) with brain mapping for maximal safe resection. We performed a retrospective study of functional and survival outcomes for left-sided eloquent GB, comparing these surgical approaches. We included 87 patients with primary left-sided eloquent GB from two centers, one performing AC and the other biopsy or resection under GA. We assessed Karnofsky performance score (KPS), language and motor deficits one month after surgery, progression-free survival (PFS) and overall survival (OS). The 87 patients had a median PFS of 8.6 months [95% CI: 7.3-11.6] and a median OS of 20.2 months [17-3-24.4], with no significant differences between the three surgical approaches. One month after surgery, functional outcomes for language were similar for all approaches, but motor function was poorer in the biopsy group than in other patients. The proportion of patients with a KPS score > 80 was higher in the resection with AC group than in the other patients at this timepoint. We detected no real benefit of a resection with AC over resection under GA for left-sided eloquent GB in terms of survival or functional outcomes for language. However, given the poorer motor function of biopsy patients, resection with AC should be proposed, when possible, to patients ineligible for surgical resection under GA, to improve functional outcomes and patient autonomy.