Tool · As of May 2026: Comet browser globally available on iOS, Android, Mac and Windows free of charge. Opus 4.6 is the default Comet Assistant model for Max users; Sonnet 4.6 for Pro. Deep Research now generates presentations, spreadsheets, dashboards and websites. Sonar API expanded into a multi-provider marketplace (OpenAI, Anthropic, Google, xAI callable through one key). Amazon won a preliminary injunction against Comet's agentic access to its site in early 2026; watch for further agentic-commerce legal action.
Perplexity
The answer engine. When you need a sourced answer, not a chat, this is the one.
Overall 8/10researchdeep_researchreal_time_searchweb_browsingautomationdocument_reviewcompliancedata_analysis
Consilix take
Perplexity is the tool you reach for when an answer with footnotes matters more than a beautifully drafted paragraph. For UK SME operators it earns its seat alongside ChatGPT and Claude precisely because it does not pretend to be a writing tool; it is a search tool with reasoning bolted on. The Comet browser is genuinely interesting (agentic browsing actually works for shopping, form-filling and inbox triage), but the legal noise around it, including the Amazon injunction in early 2026, means treat agentic commerce as experimental, not production.
Reach for it when
When the question is factual, time-sensitive, or needs to be footnoted. UK market sizing, supplier verification, competitor scans, regulatory checks, due diligence pre-reads, sourcing statistics for a proposal. Anywhere a stakeholder might ask "where did you get that number" and you need to point at a link rather than wave at the AI.
Do not reach for it when
When the deliverable is the prose itself. Claude is a far better drafter; ChatGPT is a far better generalist for image and document handling; both produce more polished long-form output. Perplexity is not the tool for writing a proposal, a contract, a board paper, or any code; it is the tool for sourcing the facts that go into them.
Core strength
Real-time, citation-backed research. Every claim links to a source you can click. Faster and more trustworthy than ChatGPT or Claude for fact retrieval, market sizing, regulatory checks, and competitor scans.
Sub-products (6)
The paid individual tier at 20 dollars per month (or 200 dollars per year). Gives unlimited Pro searches, access to a model picker covering GPT-5, Claude Opus 4.6, Sonnet 4.6, Gemini, Mistral and Sonar Pro, file and document uploads, image generation, and 5 dollars per month of API credits. The baseline product for any operator using Perplexity seriously.
Key features
Model picker across GPT-5, Claude Opus 4.6, Sonnet 4.6, Gemini, Mistral, Sonar ProUnlimited Pro searches with persistent numbered citations on every claimFile and PDF uploads with sourced analysisImage generation via DALL-E and SDXL bundled in5 dollars per month of API credits included
Best use cases
- Verify a supplier or contractor claim against public sources in under two minutes
- Pull a UK market sizing with cited ONS, IBISWorld and trade press figures
- Compare three accounting platforms with linked product pages and review sources
- Summarise a 40-page PDF tender document and surface the clauses that need legal review
- Build a competitor pricing scan with every figure footnoted
Weaknesses
Drafting and long-form writing is mediocre next to Claude. The model picker is useful but adds decision fatigue. Output formatting is functional, not polished. Cannot match ChatGPT for image understanding or Claude for code generation.
Pricing
20 dollars per month, or 200 dollars per year (saves ~17 percent). Education tier at 10 dollars per month for verified students.
Reach for this when you need a citation, not a vibe. Pair with Claude for the writing.
The premium individual tier at 200 dollars per month, launched mid-2025. Targets power users who want priority access to new features, higher Deep Research limits, choice of the strongest agent models in Comet (Opus 4.6 by default), and early access to file and app generation. Sits opposite ChatGPT Pro and Claude Max at the same price point.
Key features
Priority access to new Perplexity features (Comet model choice, file generation, dashboards)Higher Deep Research quotas and faster queueOpus 4.6 default for Comet AssistantGenerate presentations, spreadsheets, dashboards and websites directly from Deep ResearchEarly access to Labs and Model Council
Best use cases
- A consultant running multiple Deep Research reports per day for client briefings
- Heavy Comet users running agentic browser tasks who want Opus 4.6 driving the agent
- Operators who want to generate a sourced dashboard or slide deck from one Deep Research run
Weaknesses
200 dollars per month is a lot for an SME operator unless Deep Research is a daily habit. The marginal gain over Pro is real but not enormous for occasional users. ChatGPT Pro and Claude Max are stronger picks at the same price for drafting and coding.
Pricing
200 dollars per month, or 2,000 dollars per year.
Only justifiable if Deep Research is a daily tool. Otherwise stick with Pro.
Perplexity's autonomous research mode. Runs on Claude Opus 4.5 and 4.6 for Pro and Max users. Reads hundreds of sources, reasons across them, and produces a multi-page cited report in under three minutes. As of 2026 it can output the report as a presentation, spreadsheet, dashboard or static website.
Key features
Powered by Claude Opus 4.5 and 4.6 under the hoodSub-three-minute completion (vs five to thirty minutes on ChatGPT Deep Research)Generates presentations, spreadsheets, dashboards and websites from the same research runEvery figure and claim footnoted to a clickable sourceExamines market size, growth, regional variations, key players and regulatory factors as standard
Best use cases
- Produce a sourced market scan for a proposition pitch in the time it takes to make coffee
- Build a regulatory timeline for a UK sector (energy, construction, financial services) with cited statutory instruments
- Map a competitor landscape with linked sources for every claim
- Generate a board-ready sourced dashboard on a sector trend
- Pre-fill a due diligence pack on a target company with public-source evidence
Weaknesses
Source quality is only as good as the open web on that topic; thin sectors produce thin reports. Hallucinates less than ChatGPT Deep Research but still occasionally misattributes. The auto-generated dashboards and presentations look generic; treat them as drafts, not deliverables.
Pricing
Included in Pro (20 dollars per month) and Max (200 dollars per month). Limited free tier available.
The best fast research tool on the market. Faster and more sourced than ChatGPT Deep Research.
Collaborative workspaces inside Perplexity. A Space is a persistent thread with its own file uploads, custom instructions, model preference and shared access. Think of it as a Project in ChatGPT or a Project in Claude, but with web search and citations baked in. Accessible from the unified plus menu in the search bar.
Key features
Persistent context with uploaded files and custom instructionsShared access across team members (enterprise SSO supported)Model selection per SpaceWeb search and Deep Research run inside the Space contextAudit logs of AI-generated answers on enterprise plans
Best use cases
- A sales team running ongoing competitor monitoring with all materials in one Space
- Compliance team building a living regulatory tracker for a specific UK statutory instrument
- Founder running a fundraise: investor list, pitch materials, market evidence in one shared Space
- Ops director keeping a running supplier and contractor scan with source links
Weaknesses
Less polished than ChatGPT Projects for file-heavy workflows. Sharing controls are basic outside enterprise tiers. No deep integration with Google Drive or M365 the way Gemini or Copilot has.
Pricing
Included in Pro and Max. Enterprise Pro at 40 dollars per seat per month adds admin controls, SSO and audit logs.
Use when a research thread needs to live longer than one session and stay sourced.
Perplexity's AI-native web browser. Chromium-based, with an assistant pane embedded in every page that can summarise, extract, cross-reference across tabs, and act agentically (book a flight, fill a form, order groceries, triage an inbox). As of 2026 it runs on iOS, Android, macOS and Windows; the free tier is fully functional. The agent is powered by Sonnet 4.6 for Pro users and Opus 4.6 for Max.
Key features
Assistant pane on every page with summarise, extract, cross-tab reasoningAgentic task execution: form filling, booking, inbox triage, shoppingRemembers preferences across sessions for shopping and orderingPhone number actions (call, FaceTime, save to contacts) on iOSFree tier is fully functional, no subscription needed for daily use
Best use cases
- Triaging a noisy inbox without leaving the browser
- Comparing supplier websites across three tabs and producing a sourced summary
- Filling a long online form (insurance application, tender response cover sheet) from a prior PDF
- Booking travel with a remembered set of preferences
- Summarising a long-read article inline while reading it
Weaknesses
Agentic commerce is on shaky legal ground; Amazon won a preliminary injunction in 2026 blocking Comet from accessing password-protected Amazon pages. Privacy posture is less clear than Safari or Firefox; treat sensitive sessions with care. Still a young browser, occasional rough edges on extension compatibility and password manager handoff.
Pricing
Free on all platforms. Pro users get Sonnet 4.6 as the default agent model; Max users get Opus 4.6.
Genuinely useful as a second browser for research-heavy sessions. Do not put it on the critical path for anything regulated.
Perplexity's developer API. Gives programmatic access to its search-grounded LLMs. Four tiers: Sonar (lightweight factual), Sonar Pro (multi-source complex queries), Sonar Reasoning Pro (analytical with chain-of-thought), and a Deep Research endpoint. In 2026 the platform also resells OpenAI, Anthropic, Google and xAI models at provider rates plus 0.005 dollars per web search call.
Key features
Four Sonar tiers covering quick lookups to multi-step researchCitations returned as structured fields, not buried in proseNow a multi-provider marketplace: OpenAI, Anthropic, Google, xAI all callable through one keyPay-as-you-go pricing with no monthly minimumDesigned for grounded answers, not free-form generation
Best use cases
- Build an internal Q&A tool that must always cite sources
- Power a compliance assistant that needs to ground every answer in policy documents and live web sources
- Add a sourced search box to a customer-facing knowledge base
- Replace a bare LLM call in an agent workflow where hallucination is the main risk
Weaknesses
More expensive per token than calling OpenAI or Anthropic directly. The search-grounding fee adds up at scale. Latency is higher than a plain LLM call because it actually goes to the web. Documentation is decent but not as mature as OpenAI or Anthropic's.
Pricing
Sonar at 1 dollar per million input and output tokens; Sonar Pro at 3 dollars input, 15 dollars output per million tokens; Sonar Reasoning Pro at 2 dollars input, 8 dollars output. Plus 5 to 14 dollars per 1,000 requests depending on tier.
Reach for this when your product cannot afford to hallucinate and the answer must be sourced.
MCPs and connectors worth knowing (3)
Perplexity MCP serverLets Claude Desktop or Claude Code call Perplexity Sonar for grounded web search inside a Claude conversation. The cleanest way to combine Claude's writing with Perplexity's sourcing.
9/10Comet browser extensionsComet inherits the Chromium extension ecosystem; password managers (1Password, Bitwarden) and ad blockers all work. Worth knowing if you are evaluating Comet as a daily browser.
6/10Sonar API direct integrationFor n8n, Zapier, Make and custom workflows that need a grounded-search step. Cheaper than running Claude or GPT and asking them to search.
7/10Last verified: As of May 2026: Comet browser globally available on iOS, Android, Mac and Windows free of charge. Opus 4.6 is the default Comet Assistant model for Max users; Sonnet 4.6 for Pro. Deep Research now generates presentations, spreadsheets, dashboards and websites. Sonar API expanded into a multi-provider marketplace (OpenAI, Anthropic, Google, xAI callable through one key). Amazon won a preliminary injunction against Comet's agentic access to its site in early 2026; watch for further agentic-commerce legal action.