Perplexity introduced Hybrid Compute in early September, an update to its computer application that enables users to distribute AI workloads across a combination of cutting-edge models running natively and local models. In fact, the company has ultimately made it possible for users to run the platform entirely on their own devices, which seems to be just a matter of time. As long as you have a latest Apple Silicon Mac with at least 36GB of RAM, you can now run the entire computing experience on your device, including crucial controllers and coordinators.
Similar to Hybrid Compute, Perplexity views this feature as a boon for budget-conscious users and those looking to protect data privacy. For example, by accessing your Notes and Mail applications, you can ask the computer to generate a prioritized to-do list for you without sending any data from your emails to the cloud.
The company explained, “As workflows that require local execution become feasible, the machines people already own will take on more AI work.” This gives users greater control over sensitive data and cost expenditures. With today’s release, users can choose to run one of two local models: Qwen 3.6 or PPLX Pro. The latter is Perplexity’s derivative of Qwen 3.6, which the company has fine-tuned through post-training for computer use.
If at any time you feel that the task you are running would benefit from the support of cutting-edge systems, you can authorize the software to call upon one of over 15 models. Any code or tool execution required for the task will take place in an isolated sandbox environment.
Perplexity Computer Makes Local AI Workloads a Reality
Perplexity stated, “The same security protocols persist in the cloud version of Perplexity Computer.” Perplexity Pro, Max, and Enterprise subscription users can start using the computer locally from today. Once you download and install the latest version of the Perplexity Mac application, simply follow the in-app instructions to download and install the local models.
Item Specification RAM 36GB