Knowledge catalogue
“AI Meltdown” is the scenario where the agent goes off the rails and forgets/ignores all previous instructions and soft guardrails. No prompt injection or malicious actor is needed for this to happen.
Today we’re open-sourcing Numbat, an agent-detection and response layer that is designed to work across agent harnesses. Numbat gives security teams visibility into agent activity, with controls to bl
Introducing SPACE, the sandbox platform behind Perplexity Computer. It creates isolated environments for code, files, and long-running agent sessions. SPACE has handled 100% of Computer production tra
WANDR tests how well agents discover large sets of entities and verify specific facts about each one. It provides a dense, interpretable eval signal that reveals whether an agent fails, and where. The
We’re open sourcing WANDR. WANDR is an internal benchmark we built and used for building deep and wide research capabilities inside Perplexity Computer. https://research.perplexity.ai/articles/wandr-b
Perplexity has released two new AI models, GPT-5.6 Terra and Sol, which are now available for use in Perplexity's search platform and Perplexity Computer tool. These models represent updates to Perple
Grok 4.5 is now available as an orchestrator model in Computer for Consumer Pro and Max subscribers. We evaluated it against five other orchestrator configurations on WANDR. It scored higher than ever
Perplexity AI has launched a new analytics feature that allows users to track their credit spending across different AI models. The feature is accessible through the Analytics section in Account Setti
The model has an advisor tool that natively escalates to a stronger model when needed. This model is hosted in the U.S. by Perplexity on Nvidia B200 GPUs. We will improve the model in research preview
Perplexity conducted a cost efficiency comparison of language models, using GLM 5.2 as the baseline metric for cost per task. Results showed GLM 5.2 with an advisor achieved 2.1x cost efficiency on th
We're releasing a research preview of a new orchestrator model in Perplexity Computer. The model is an adapted version of GLM 5.2, post-trained for the Computer harness. It delivers near-frontier perf
Claude Fable 5 has been re-enabled as an orchestrator model option within Computer, Anthropic's AI system for automated task execution. This announcement from Perplexity indicates that users can again
Claude Sonnet 5 is now available for use in Perplexity, accessible to Pro and Max tier subscribers. Users can also select Claude Sonnet 5 as an orchestrator model within Perplexity's Computer feature,
Introducing Computer for Counsel. Computer now connects the research databases, document tools, and matter-management systems lawyers use every day. Pull citable sources from @midpageAI, @LegalZoom, @
Computer for Counsel is an enterprise offering from Perplexity AI designed specifically for legal professionals and law firms to leverage AI-powered research and analysis capabilities. The solution li
Claude Fable 5 is now available in Computer as an orchestrator model designed by Anthropic for handling long and complex tasks. This represents Anthropic's latest state-of-the-art model offering enhan
The Billion Pound Build is a Perplexity-hosted initiative inviting users to submit project pitches, likely focused on generating innovative ideas or solutions. This appears to be a community engagemen
Yesterday at London Tech Week we announced the Billion Pound Build competition. Teams can enter to secure a share of £1M in Computer credits by using Perplexity Computer to build their company. The pi
1,500 teams entered the Billion Dollar Build and have been building for 7 weeks. 8 finalists present live on June 9 to our judges @LewisHamilton (7x World Champion Driver), @AravSrinivas (CEO, Perplex
A study or analysis demonstrates that computer technology enhances user autonomy and task quality while reducing both time and cost requirements, enabling users to undertake a broader range of tasks t
Research indicates that autonomous agents with greater operational autonomy, particularly in computer-based tasks, demonstrate improved performance quality and user satisfaction outcomes. The study su
We published new research with Harvard on the shift from chat interfaces to autonomous agents like Computer. Over 3 months, findings show workers using Computer finish tasks in 87% less time at 94% lo
Nvidia's Nemotron 3 Ultra, an open-source model designed for long-context tasks, is now available to Pro and Max subscribers on Perplexity and Computer. This new model represents Nvidia's latest contr
Today we're launching the Main Street AI Accelerator with the U.S. Small Business Administration. We're committing 25M in Computer credits, 250 each for up to 100,000 eligible companies, in honor of A
Perplexity is inviting users to join a waitlist for a Windows version of its Personal Computer product, which appears to be an AI-powered tool or application designed to run on Windows operating syste
Perplexity Computer is for growing businesses. Computer connects to 400+ tools for every type of company, including Intuit QuickBooks, Vercel, Shopify, Canva, and more. Learn more about how people are
Personal Computer is coming to Windows. Personal Computer for Windows runs on your machine and orchestrates across the apps and files you use every day. We'll roll out first to paying Max and Enterpri
Perplexity AI demonstrates a hybrid approach to AI inference that combines on-device processing for sensitive data with cloud-based AI services to achieve scalability and contextual understanding. Thi
Perplexity discusses hybrid agentic inference in Perplexity Computer, which likely describes a computational approach that combines processing between local machines and data centers. The concept sugg
Today we're announcing that hybrid agentic inference is coming to Perplexity Computer. Computer can split tasks between a local model running on your machine and frontier models in the cloud. This kee
Two new ways to bring your health data into Perplexity. Perplexity now connects to Apple Health on iPhone. Use your sleep, activity, and HRV data in Computer. Function is now available in Perplexity H
Introducing Search as Code, our new search architecture for AI agents. It writes Python that calls our search stack directly, instead of looping through function calls one at a time. Available in the
The Perplexity Agent API documentation includes a 'Search as Code' feature accessible through the sandbox tools section, enabling developers to integrate search functionality programmatically within a
WANDR is our in-house wide benchmark, built to mirror real professional research workloads. Search as Code scores 0.386 to the next best system's 0.152, and the benchmark is far from saturated. We're
Claude Opus 4.8 has been made available to Max subscribers on the Perplexity platform and Computer application. This release expands access to Anthropic's Claude model through Perplexity's subscriptio
Computer runs on the same secure infrastructure as the rest of Perplexity. The enterprise-grade security layer includes SAML SSO, audit logs, and granular admin controls. Read more: https://www.perple
Perplexity Computer is now available inside Microsoft Excel, Word, PowerPoint, and Outlook. Orchestrate across work with Computer directly in the side panel of your app to draft documents, model, buil
At production input lengths, the encoder cuts p50 latency by roughly 5× vs. HuggingFace tokenizers, 2× vs. SentencePiece C++, and 1.5× vs. IREE C. At 514 tokens, it runs in 63 µs with zero heap alloca
Perplexity published a blog article detailing methods to improve the CPU performance of Unigram tokenizers, addressing computational efficiency in natural language processing. The article explores opt
We're open-sourcing the Unigram tokenizer we rebuilt to reduce CPU utilization by 5-6x. Small rerankers and embedders run in single-digit milliseconds on GPU, making CPU tokenization a meaningful shar
This post demonstrates how Perplexity AI can assist e-commerce store owners by automating multiple tasks simultaneously, including market research, AI-generated product imagery, and store design. The
Bumblebee started as an internal tool. Making Perplexity products more secure for users starts with protecting the developer systems we use to build them. Read the full blog: https://www.perplexity.ai
Today we're open-sourcing Bumblebee, a read-only scanner for macOS and Linux. It checks developer machines for risky packages, extensions, and AI tool configs. Connected to Computer, it can trigger de
Context compression isn't new in RAG. Our contribution is making it query-aware, citation-preserving, and fast enough for orchestration. Read the full research blog: https://research.perplexity.ai/art
Less noise = more signal. Vital content per snippet is up 63%. Ads, navigation, metadata, and unhelpful content are culled before handoff to the answer model. On SimpleQA, we achieve a 50x compression
We've productionized query-aware compression for faster, cleaner, more-accurate search. Better context is better than more context. Our system cuts context tokens up to 70% while improving answer qual
Rho cut weekly meeting time by 90% with Perplexity Computer. Computer checks Slack, Notion, Jira, Figma, and Google Docs, then flags missing tasks and changes the team needs to see. 120 work hours sav
Build dashboards and automations from your Snowflake data for pipeline analysis, product usage, customer segments, and more. Admins maintain control over access, business definitions, and shared data
Computer now connects to Snowflake. Run end-to-end work against live warehouse data and get answers with SQL, source tables, filters, and metrics. It’s like a personal data science team, on call with
Computer is secure by default. Every task runs in its own hardware-isolated sandbox with VPC-level storage and compute separation. Agents are authenticated with short-lived proxy tokens instead of raw
External content is scanned in parallel by ML classifiers and the BrowseSafe model before agents act on it. File connector data is encrypted in transit and at rest, uploaded files automatically delete
PayPal runs 74,000 weekly tasks in Perplexity Enterprise. Teams use it for model validation, channel performance, market trend research, competitive intelligence, and product analysis. Read the custom
The benchmarks show the gap. NVLS all-reduce latency drops from 586.1µs on H200 to 313.3µs on GB200. In MoE prefill at EP=4, combine falls from 730.1µs to 438.5µs. For decode, GB200 sustains much high
This NVIDIA remains the strongest platform for large-model inference at scale. Prefill/decode disaggregation, Blackwell-native quantization, custom kernels, and rack-scale NVLink turn GB200 into faste
We published new research on how we serve post-trained Qwen3 235B models on NVIDIA GB200 NVL72 Blackwell racks. GB200 is a major step up over Hopper for high-throughput inference on large MoE models,
We've published our internal manual for building agent skills. Skills require a new way of thinking for developers. https://research.perplexity.ai/articles/designing-refining-and-maintaining-agent-ski
This post draws a parallel between Python's design philosophy (from 'The Zen of Python') and machine learning model capabilities, suggesting that if a skill or implementation can be easily explained,
This post contrasts two philosophical approaches: Python's 'Zen of Python' principle that special cases shouldn't violate general rules, versus a 'Zen of Skills' perspective arguing that gotchas and e
Perplexity has released a new Mac application for its Personal Computer feature, which users can download from perplexity.ai/personal-computer. The company is discontinuing support for the older versi
Personal Computer is built for the Mac ecosystem, so you can manage it from anywhere. Start tasks from your iPhone with local files on your Mac. Add a Mac mini for full continuity and always-on agents