Perplexity said on 20 October 2026 that Perplexity Computer, its AI agent, can now run fully local on Macs with Apple silicon, “making it possible to run the AI model, harness, and task execution entirely on device,” according to a post on its blog. “Local tasks use PPLX Pro, our post-trained version of Qwen 3.8, and don’t consume Computer credits,” the company said on X.
The local mode is available to Pro, Max and Enterprise subscribers in the Perplexity Mac app, and Perplexity recommends 36 GB or more of memory. Below: what runs where, what it means for credits, and how it extends Perplexity’s earlier local releases.
Business Pill · COST PER TASK
A short explainer on cost per task, paying for the job rather than the word. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, Perplexity now offers three places to run the same agent: fully on a Mac, split between the Mac and the cloud, or in the cloud, and it says local tasks don’t consume Computer credits.
“When given permission, Computer can work across local files and Mac apps like Notes and Mail without sending personal data to the cloud,” the blog post says. Perplexity wrote that this “gives users more control over sensitive data and credit spend.”
Computer “runs Qwen 3.6 or PPLX Pro, a post-trained version of Qwen,” according to the blog, which says it “launches with Qwen 3.6 and PPLX Pro as local model options.” Perplexity’s X post calls PPLX Pro “our post-trained version of Qwen 3.8.”
Perplexity’s blog says Computer launches with Qwen 3.6 and PPLX Pro as local model options and that the local orchestrator can call on one of 15+ frontier models; Perplexity said on X that local tasks don’t consume Computer credits. Circles are sized by area.
Fully Local Versus Hybrid
The Mac app offers two ways to run Computer, the blog says: “Hybrid Compute starts in the cloud, coordinating research and reasoning with local work on private files. Running Computer locally starts on the Mac, keeping the task’s planning, reasoning, and file processing on device.”
Users “can authorize Computer to escalate to the cloud for current information, browser use, and connected apps,” and “the local orchestrator can call on one of 15+ frontier models for advanced reasoning,” according to the post. In Perplexity’s examples, a portfolio analysis on private holdings can run entirely on device, while “a deal review that combines current market research with confidential financials can use Hybrid Compute.”
Where It Runs and What It Touches
Supported apps include Notes, Messages, Mail and Numbers, the blog says, and “code and tool execution run in isolated sandbox environments with controlled access to files and connected apps.” Perplexity names “the new Mac mini with M6 and M5 Pro” as examples of supported Apple silicon Macs.
To start, the post says, users download the latest Perplexity Mac app, follow the in-app instructions to set up a local model, and choose fully local to run the task on the Mac.
Connecting the Coverage
Perplexity says it first brought Computer to local files and native Mac apps with Personal Computer for Mac in May, and added Hybrid Compute last month. In August we covered Perplexity and Nvidia’s Portable Computer, which launched on Linux with local work consuming zero billing credits.
As we read it, the credit line is the business detail: Perplexity’s blog says running locally gives users “more control over sensitive data and credit spend,” so a task’s cost depends on where the user chooses to run it.
As we read it, the cloud stays one permission away: the post says users can authorize escalation for current information, browser use and connected apps, and that the local orchestrator can call on one of 15+ frontier models.
As we read it, this is the third local step this year in Perplexity’s own telling, after Personal Computer for Mac in May and Hybrid Compute in September, and it follows the Linux-based Portable Computer we covered in August.
Perplexity blog, 20 October 2026
“With both the model and agent running locally, sensitive work can stay on the Mac through every stage of a task, from planning and analysis to the finished output”
Three Implications
NO CREDITS FOR LOCAL TASKS Perplexity said on X that local tasks use PPLX Pro and don’t consume Computer credits.
36 GB OR MORE Perplexity recommends 36 GB or more for running Computer locally on supported Apple silicon Macs.
PRO, MAX AND ENTERPRISE The local mode is available to Perplexity Pro, Max and Enterprise subscribers in the Perplexity Mac app, the blog says.
The framework puts it this way: “Inference is a revenue engine: continuous, distributed across millions of applications and agents, incurring marginal cost with every query.”
As we read it, a fully local Computer task runs that marginal cost on the user’s own Mac, and Perplexity says such tasks do not consume Computer credits. The framework was published in March 2026 and does not discuss Perplexity’s local mode.
What Is Not Established
The blog and X posts we read give no figures on how many users run Computer locally, how local task speed compares with the cloud, or what a cloud task costs in credits. The 36 GB figure is a recommendation, not a stated minimum, and the two Perplexity posts describe the Qwen base under PPLX Pro in different words.
Business Engineer Framework
The State of The Inference Economy
A Business Engineer read on how inference became the revenue engine of AI, and who pays the marginal cost of every query.
Perplexity says Computer can now run fully local on Apple silicon Macs, with local tasks not consuming Computer credits and cloud models available when users authorize them. As we read it, that puts more of the agent’s running cost on hardware the user already owns, which Perplexity itself frames as more control over “sensitive data and credit spend.”
A note on sourcing. This piece is built on Perplexity’s 20 October blog post and product page, read in full, its three-post X thread of the same day, and our own earlier pieces on Perplexity Computer. We did not test the local mode ourselves and did not contact Perplexity. The chart’s circles are sized from the counts Perplexity gives. Nothing here is a product endorsement.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
