All tasks / Tools
Perplexity
Search-based answer engine on web, desktop and mobile that cites sources; paid plans can pick third-party models.
- Perplexity AI
- Cloud
- Free tier · from $20/mo
Where it ranks
- SChat & assistantsApps + Runtimes & gateways · Cloud · via Gemini 4 Argon
- BSearch & researchApps + Search APIs & engines · Cloud · 43 on DeepResearch Bench (RACE, GPT-5.5 judge)
- SImage understandingApps + Runtimes & gateways · Cloud · via GPT-6.1 Sol
- –Browser & computer agentsAgent apps & AI browsers · Cloud · no rated model available here yet
Ranked by its own benchmark result where it has one; otherwise by the best model it offers.
Facts
- Vendor
- Perplexity AI
- Approach
- Apps, Runtimes & gateways, Search APIs & engines, Agent apps & AI browsers
- Where it runs
- Cloud only: needs the vendor’s service
- Interface
- Web app, Desktop app, Mobile app
- Skill level
- Beginner: works out of the box
- Price
- Free tier · from $20/mo
- Pricing notes
- Free: automatic 'Best' model and limited Pro Searches; Pro $20/mo ($200/yr) selects GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 3.8 Flash, Grok 4.7, Kimi K3, GLM 5.3, Nemotron; Max $200/mo adds GPT-6 Astra, Gemini 4 Argon, Claude Opus 5.5 and Fable 5.1.
- Source code
- Closed source
- Last verified
- 18 Oct 2026
Price from perplexity.ai, 18 Oct 2026.
Models it gives you
- SGemini 4 ArgonGoogle · closed
- SClaude Opus 5.5Anthropic · closed
- SClaude Fable 5.1Anthropic · closed
- SGemini 3.8 FlashGoogle · closed
- SKimi K3Moonshot AI · open weights
- SGPT-6.1 SolOpenAI · closed
- AGLM-5.3Z.ai · open weights
- AGPT-6 AstraOpenAI · closed
- AClaude Sonnet 5.5Anthropic · closed
- AGrok 4.7xAI · closed
Its own benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| DeepResearch Bench (RACE, GPT-5.5 judge) | 43 | perplexity-ResearchRACE overall score on the maintained GPT-5.5-judge leaderboard (data_gpt55/leaderboard.csv, space last updated 2026-10-11); rank 16 of 17; comprehensiveness 41.78, insight 41.27, instruction following 45.31, readability 46.03. Perplexity's Research mode as run for the 2025 DRB paper; the product has since changed. | 11 Oct 2026 | huggingface.co |
| DeepResearch Bench II (GPT-5.5 text-only eval) | 33.6% | Perplexity Research2025 product run (DRB2 paper), re-judged with GPT-5.5; reflects the 2025 product. | 19 Oct 2026 | huggingface.co |
| Mind2Web 2 (agentic search, private test set) | 15% | Perplexity Deep ResearchMind2Web 2 paper run (mid-2025); reflects the 2025 product. | 19 Oct 2026 | osu-nlp-group.github.io |
Compare
Sources
- https://www.perplexity.ai/help-center/en/articles/10354919-what-advanced-ai-models-are-included-in-my-subscription
- https://www.perplexity.ai/help-center/en/articles/11187416-which-perplexity-subscription-plan-is-right-for-you
- https://www.perplexity.ai/help-center/en/articles/11680686-perplexity-max
- https://www.perplexity.ai/hub
- https://www.semafor.com/article/02/25/2026/perplexity-launches-computer-super-agent
- https://huggingface.co/spaces/muset-ai/DeepResearch-Bench-Leaderboard
- https://www.computerworld.com/article/4138723/perplexitys-new-computer-agent-will-run-other-agents-for-you.html
- https://itdaily.com/news/cloud/perplexity-ai-launches-perplexity-computer/
- https://docs.perplexity.ai
- https://osu-nlp-group.github.io/Mind2Web-2/