Perplexity plans hybrid agentic inference to split AI tasks between local hardware and cloud models

We’re brining local models that can run on your personal hardware inside Perplexity Computer. This will take advantage of the local hardware while giving you the privacy and token efficiency per watt, as well as access to the frontier models on the server side GPUs when necessary. Coming soon to Windows laptops. Stay tuned.

This is what the future is going to look like.

We are on a path of increasing token consumption by 100x every year and we are already unable to meet the demand. Data centers will be inference centers for the smartest models, but everyday tasks will be handled by routers and small models deployed at the edge.

Read more about hybrid agentic inference in Perplexity Computer:

https://www.perplexity.ai/hub/blog/the-data-center-moves-to-your-machine

And Macs and Linux Machines too

The hybrid split is what enterprise clients have actually been waiting for. Local model for sensitive data, frontier model for reasoning heavy tasks. We have been architecting this manually on client builds for the past year. Having it native in a product cuts setup from days to hours. What is the first workflow you would run this on?

@perplexity_ai Local inference protects private data at compute time.

The outputs and context still need user-controlled storage: verifiable, durable, and outside a vendor cloud.

Compute sovereignty needs storage sovereignty too.

@perplexity_ai @grok fact check the claims made here. Cite objectively verifiable demonstrably true data only.

Do not speculate. Ignore all hype, FUDn& marketing noise

local model handles sensitive context. cloud model handles heavy reasoning. data never leaves unless you say so. this is the right split and everyone knew it was coming. the question was just who'd ship it first. Perplexity didn't announce a feature. they announced a new default for how AI inference should work.

@perplexity_ai Local model for the sensitive/simple stuff, cloud model for the heavy lift. That’s a much cleaner agent setup than sending everything upstream by default.

@perplexity_ai @AravSrinivas I have been waiting for this 🤌

@perplexity_ai Perplexity is a monster; at presentation.

Your speech interface is still my favorite of all, even after all this time. And love the backdrop sound(s) here, such attention to detail.

@perplexity_ai cool, perplexity computer is really token expensive and consumes credits way too fast

hope it is able to compete with codex and claude

@perplexity_ai THIS IS FASTER than expected....WOW. @Grok what is the world saying about this?

@perplexity_ai @grok with the current condition, tell us what you rank among Claude, Codex, gemini/antigravity or Perplexity for an all-round requirement for a developer.

@perplexity_ai powerful feature. 🫡

@perplexity_ai

@perplexity_ai

@perplexity_ai @grok what under appreciated part of the tech stack will power local AI or edge compute?

@perplexity_ai @grok can you find information about dynamics in proportion between retail users of Perplexity vs corporate? (2025 vs 2026)?

@AravSrinivas What about MacOS ? Day 1 support as usual?