Perplexity Computer now works in your own Chrome
A new extension links the Mac and Windows apps to sites you already use, and Comet still does not need it.
Every Perplexity update in one feed — release notes, features and changes, straight from the source.
87 updates
A new extension links the Mac and Windows apps to sites you already use, and Comet still does not need it.
One voice session can span several Computer sessions, start tasks, and manage Projects. It is on the Mac app and the web for Pro and Max.
HeyGen’s HyperFrames skill, now in Perplexity Computer’s skills marketplace, turns a written description of a video into a finished MP4.
The Impeccable skill, now in Perplexity Computer’s skills marketplace, reviews an interface for layout, type, contrast and accessibility and fixes the issues.
The Taste Skill in Perplexity Computer’s skills marketplace sets an art direction before Computer builds a site, drawing on 247 award-showcase websites.
The new skill scans PHP sites and GitHub deployment configs for exposure and ranks the fixes. Runs are free through October 31.
Both the model and the agent now run on supported Apple silicon Macs, so Computer can work with your files and Mac apps without sending data to the cloud.
Inspect a 3D preview, export editable CAD and review FreeCAD edits through a video of the changes.
Version 26.37.1 supports parallel work across tabs, windows and screens for Computer Mac users.
The vendor’s flight-game example used 97 credits; that is an example cost rather than a fixed price.
Search calls on the affected presets now cost $1 per 1,000 invocations, plus model tokens, and run about 800 ms faster.
Any Perplexity API key works, with input priced at $0.04 per million tokens, free output tokens and no per-request fee.
Developers can send five document formats inline or by public HTTPS URL, then ask follow-ups about inline files without uploading again.
Developers can filter sources and formats in Agent API requests, with successful searches costing $2.50 per 1,000 calls.
Successful sandbox extraction calls now count towards Agent API response totals, with pricing unchanged and no historical backfill.
Requests with up to 100,000 input tokens cost $0.10 per million input tokens, with higher rates above that threshold.
Input tokens now cost $0.02 per million, and requests using the previous model name continue to work at the lower rate.
The Agent API adds a text-only reasoning model with a 1M-token context window, priced at $0.20 per million input tokens, replacing Nemotron 3 Ultra.
The 27B decider stays at $0.02 per million input tokens, and a new 4B flash model costs $0.01, with both accepting text and images.
Gemini 4 Argon costs $2.00 per million input tokens and $10.00 per million output tokens through November 30, 2026, on default processing only.
Perplexity's Agent API docs drop old Claude, Gemini and Grok models, and list GPT-6 Astra at $10 per million input tokens and DeepSeek V4 Pro at $1.32.
Developers can set require_approval to "always" or a filter, then approve or deny each call; Python SDK 0.43.8 and TypeScript SDK 0.38.8 add the types.
The Agent API preset now allows 512,000 output tokens, and frozen configurations need max_output_tokens updated to match.
Agent API web_search calls under the low to xhigh presets now cost $1.00 per 1,000, down from $2.50, and run about 800 ms faster.
Perplexity Computer can now pull DocSend deal room files to keep documents like investment theses current.
GPT-6 Sol is now Perplexity Computer's default Light-effort model, and is also available in Perplexity itself.
Perplexity adds effort levels to Computer, opens a Skills Marketplace, and brings GPT-6 Astra to Pro and Max plans.
Perplexity Computer adds AI video generation via MiniMax H3 and Seedance 2.5, available to Pro and Max subscribers.
The Windows feature keeps file and app work local on RTX hardware, while still allowing frontier cloud models when needed.
Perplexity Computer adds @ mentions on the web, letting users tag and share sessions with coworkers.