AI Assistant Summary: Perplexity made 2026 its “browser year.” The flagship release was the Comet browser, which launched globally to all iOS users (joining Android, Mac, and Windows). Comet’s agentic browsing capabilities are now integrated into Samsung Internet, and Comet for Enterprise is deployable via MDM. Beyond the browser, Perplexity upgraded Deep Research to state-of-the-art performance, now running on Claude Opus 4.5 for Max and Pro users, with the ability to generate presentations, spreadsheets, dashboards, and websites directly. Voice Mode now uses OpenAI’s GPT Realtime 1.5 (25%+ more reliable). Max subscribers can pick the model powering their browser agent (Opus 4.6 default, Sonnet 4.5 alternative). This guide summarizes every major announcement with direct links to Perplexity’s official changelog. Updated through May 14, 2026.
Perplexity in 2026 looks fundamentally different from Perplexity in 2025. The product expanded from “answer engine” to a full browser (Comet), a deep-research agent, and a multi-platform AI ecosystem. This guide compiles the consequential 2026 announcements with direct links to Perplexity’s official changelog and blog.
The headline: Comet browser, globally available
Comet is Perplexity’s AI-native browser — their attempt to redefine what a browser is when AI is built into every interaction. Started in 2025 as an invite-only beta. In 2026, Perplexity rolled it out:
- Global availability: Available now to everyone worldwide, across iOS, Android, Mac, and Windows.
- Native AI in every browsing action: Summarize any page, ask follow-up questions, have Comet do tasks in your browser autonomously.
- Comet Assistant for in-page research and summarization.
- Source: perplexity.ai/hub/blog/comet-is-now-available-to-everyone-worldwide
Comet competes with: Arc (now Atlas Browser), Brave with AI features, Microsoft Edge with Copilot, and the upcoming OpenAI browser project. Perplexity’s distinctive angle: their answer-engine background makes Comet best-in-class for citation-grounded research while browsing.
1-on-1 Coaching
Claude AI Crash Course
1-hour private video session with James. We cover where Perplexity fits in your research workflow (especially how it pairs with Claude for synthesis), and when Comet is worth replacing your daily browser with. Most users get the best results from combining Perplexity for research + Claude for writing.
Group Format
AI Workshops for Research Teams
Team-format workshops for research teams, journalists, analysts, and any group that does heavy desk research. We cover Perplexity + Claude + Notebook LM workflows. Best for 3+ people.
Comet for Enterprise
Comet is now available for Enterprise organizations. Enterprise administrators can deploy Comet across macOS and Windows devices via MDM (mobile device management), with central control over policies, data handling, and feature access. Includes Comet Assistant for in-page research and summarization built specifically for enterprise compliance needs.
Use cases: financial services analyst desks, legal research firms, journalism organizations — anywhere a team does heavy desk research and could benefit from Comet’s assistance without the consumer-tier privacy concerns.
Deep Research, dramatically upgraded
Perplexity upgraded Deep Research to achieve state-of-the-art performance on leading external benchmarks. Key changes:
- Now runs on Claude Opus 4.5 (Max and Pro tiers). Anthropic’s flagship for the most demanding research work.
- Structured outputs: Deep Research can now create presentations, spreadsheets, dashboards, websites, and other structured outputs directly within the product. Just prompt for the deliverable: “Create a presentation analyzing the AI agent market.”
- Performance: State-of-the-art on leading external benchmarks — including those that measure synthesis quality and citation accuracy.
This positions Perplexity Deep Research as a real competitor to OpenAI’s Deep Research and Anthropic-native workflows. The structured-output capability is the differentiator: most research agents return long text; Deep Research returns the actual deliverable.
Our Perplexity for Research guide covers the practical workflow.
Voice Mode upgraded (GPT Realtime 1.5)
Perplexity’s Voice Mode now runs on OpenAI’s GPT Realtime 1.5 with over 25% more reliable interactions and dramatically improved voice expressiveness. Rolled out to everyone on Comet for desktop and Android.
Compare to: ChatGPT’s Advanced Voice, Gemini’s voice mode, ElevenLabs Conversational AI. Perplexity’s voice mode is differentiated by being citation-grounded — it returns sourced answers rather than free-form responses.
Choose your browser agent model
For Max subscribers, Perplexity now lets you choose the model powering their browser agent:
- Opus 4.6 (now the default) — Anthropic’s flagship through 4.6 (Opus 4.7 is available where compatible)
- Sonnet 4.5 — faster, cheaper alternative
This is unusual for an AI product: most companies lock you to their own model. Perplexity’s positioning has long been “use the best model for the job,” and the model picker for paid users underlines that.
Samsung Internet partnership
Comet’s agentic browsing capabilities have been integrated into Samsung Internet, with Perplexity available as an optional default search engine on Samsung devices. Distribution win for Perplexity: Samsung Internet is preinstalled on hundreds of millions of devices.
The unified + menu
Perplexity’s UI now has a unified + menu in the search bar bringing together:
- Deep Research — the long-form agentic research mode
- Model Council — ask multiple frontier models the same question and compare answers
- Create files and apps — build structured outputs directly
- Learn step by step — education-focused tutor mode
- Spaces (collaborative workspaces) and other existing features
The + menu makes Perplexity less “search engine” and more “AI workspace.”
How Perplexity compares to other research/search AI
| Feature | Perplexity | ChatGPT | Claude | Gemini |
|---|---|---|---|---|
| Citation-grounded answers | Yes (best) | With search | With tool use | Yes |
| AI-native browser | Yes (Comet) | Coming | No | No |
| Deep Research mode | Yes (SOTA) | Yes | Via tool use | Yes (Max) |
| Voice mode | Yes | Yes (Advanced) | No | Yes |
| Pick model (Opus/Sonnet) | Yes (Max) | No | Own only | Own only |
| Structured outputs from research | Yes | Limited | With prompting | Yes |
Perplexity’s distinctive advantage: citation-grounded, multi-model, and AI-native browser. Less differentiated as a pure chat assistant (Claude and ChatGPT still lead). Best in class for desk research and source-tied work.
Pricing reminders (May 2026)
- Perplexity Free: Limited daily Pro Searches and standard search
- Perplexity Pro: $20/month — unlimited Pro Search, Deep Research access, file uploads
- Perplexity Max: Higher tier with model picker, prioritized access, expanded Deep Research
- Comet: Free to install; some features require Pro/Max
- Enterprise: Custom pricing through Perplexity sales
Verify current pricing at perplexity.ai/pro.
Frequently asked questions
Is Comet really worth replacing my browser?
For research-heavy users: probably yes. For everyone else: install alongside your current browser and use Comet specifically for research-and-deliverable tasks. Don’t force a daily-driver switch — the AI features are the value, not the browser itself.
Should I use Perplexity instead of Claude or ChatGPT?
They solve different problems. Perplexity is a research engine with citations. Claude and ChatGPT are general assistants better at writing, reasoning, and tasks where citation grounding matters less. For most users, use Perplexity FOR finding and verifying sources, then use Claude or ChatGPT FOR synthesis and writing.
Is Deep Research better than OpenAI’s or Anthropic’s?
Perplexity claims SOTA on leading benchmarks. Independent tests show all three competitive on most research questions, with each excelling at different question types. Perplexity’s structured-output edge (returning a presentation or spreadsheet rather than just text) is a real differentiator.
Why does Perplexity let me pick the Claude model?
Perplexity’s positioning is “use the best model for the job” rather than lock-in. They route to OpenAI for voice, to Anthropic for browser agents, to their own infrastructure for fast searches. The Max model picker exposes this to users who want control.
Where do I find Perplexity’s changelog?
Two main sources: perplexity.ai/changelog (regularly updated “What We Shipped” entries) and perplexity.ai/hub (longer-form blog posts).
Is Voice Mode free?
Available to everyone on Comet for desktop and Android. The Voice Mode using OpenAI’s GPT Realtime 1.5 is now the default. Quality and reliability noticeably better than the previous version.
Can my company deploy Comet to all our devices?
Yes — Comet for Enterprise lets administrators deploy via MDM across macOS and Windows. Contact Perplexity Enterprise sales for the rollout process.
Get each Perplexity ship in plain English
Each Perplexity release, unpacked in 90 seconds
Free daily newsletter — one story, one tool, one tip. Plain English, no jargon. One issue per day.
Free forever. Unsubscribe anytime.
You may also like
- What’s New in ChatGPT 2026
- What’s New in Claude 2026
- What’s New in Gemini 2026
- What’s New in Grok 2026
- What’s New in Mistral 2026
- How to Use Perplexity for Research
- Perplexity vs Google
- Perplexity vs ChatGPT vs Claude
Sources (all Perplexity official)
- Perplexity. “The Internet is Better on Comet.” perplexity.ai/hub/blog/comet-is-now-available-to-everyone-worldwide (verified May 14, 2026)
- Perplexity. “Introducing Comet.” perplexity.ai/hub/blog/introducing-comet
- Perplexity. “Comet Browser landing page.” perplexity.ai/comet
- Perplexity Changelog. perplexity.ai/changelog
- Perplexity Pro pricing. perplexity.ai/pro
- Perplexity Hub blog. perplexity.ai/hub
Two ways to go further
The AI Prompt Library
1,000+ ready-to-use prompts for Claude, ChatGPT, and Gemini. Stop staring at a blank box.
Get it for $39 →2-Hour Live AI Crash Course
A private, beginner-friendly session across Claude, ChatGPT, Gemini, and the wider landscape.
Book for $125 →