Models.dev records 200,451 monthly visits with -3.9% month-over-month growth and Free pricing. Perplexity records 88,310,223 monthly visits with -3.7% growth and Freemium pricing.
It is a reference catalogue, not an inference or hosting service
–Data quality depends on community and maintainer updates keeping pace with releases
–Interface is aimed at developers and offers little guidance for non-technical users
Perplexity
Real-time AI answer engine and research workspace combining live web citations with frontier reasoning models.
Deep Research Autonomous Reports
Executes multi-step iterative web and document research across hundreds of sources, cross-checking facts and generating structured, citation-backed analytical reports that can be exported as PDFs or shared as interactive Perplexity Pages.
Multi-Model Routing & Model Council
Lets Pro and Max subscribers switch dynamically between Perplexity's in-house Sonar models and top external models (including Claude 4, o3-pro, and Gemini) or compare multi-model reasoning outputs side by side.
Perplexity Spaces & Hybrid RAG
Provides dedicated project workspaces where users upload local PDFs, spreadsheets, and text files, configure custom assistant instructions, and run queries that combine private document retrieval with live internet search.
Perplexity Computer & Comet Assistant
Extends Perplexity from passive Q&A into agentic execution, turning high-level user requests into completed workflows, multi-tab browser actions, and automated data synthesis via the Comet ecosystem.
Strengths & Pros
+Every generated statement links directly to numbered, verifiable web or document sources, drastically reducing ungrounded LLM hallucinations.
+Deep Research mode delivers comprehensive, multi-page synthesis reports at a fraction of the time required for manual desk research.
+Broad model optionality lets subscribers use Sonar, Claude, OpenAI, and Gemini models under a single $20/month subscription.
+Focused search scopes (Academic, Finance, Web, Social) and Spaces allow clean separation between general browsing and domain-specific knowledge work.
Limitations
–The Free tier is tightly restricted to 3 Pro Searches per day and just 1 Deep Research query per month with no manual model selection.
–Power users requiring heavy agentic Computer runs or uncapped o3-pro/Opus-level research must jump from $20/month Pro to the steep $200/month Max tier.
–Synthesized summaries can still inherit inaccuracies if underlying top-ranked web sources contain outdated or biased reporting.
#1
