Tabnine goes hybrid, serving AI models on both cloud and local

Tabnine has introduced a new Hybrid model that leverages both cloud and local servers to enhance code prediction experiences for users. This model, enabled by default for new installations as of June 1, 2022, combines the strengths of cloud-based models, which offer GPU-accelerated and more accurate code completions, with the independence of local models that run without network access. Users maintain control over their preferences and can switch between cloud, local, or hybrid modes via the Tabnine Hub. Despite the integration of cloud services, Tabnine emphasizes its commitment to user privacy by ensuring all communications are strongly encrypted and no user code is stored or shared. The introduction of this hybrid model marks a step forward in Tabnine's ongoing efforts to provide increasingly powerful coding tools.