Content History
Every millisecond matters. We’re open sourcing the tokenizer we built and deployed on production; that’s far efficient than huggingface and sentencepiece.
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 share of total latency. http://github.com/perplexityai/pplx-garden
RT Perplexity 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 share of total latency. http://github.com/perplexityai/pplx-garden
RT Brian Roemmele We have been using the open source tool Bumblebee at The Zero-Human Company over the last day and find it amazing!
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 deeper scans whenever a new supply-chain risk emerges. https://github.com/perplexityai/bumblebee
RT Vaibhav Sisinty Perplexity just open-sourced the tool they use internally to keep their own developers safe. 😨 It's called Bumblebee. It runs quietly on a developer's laptop and checks for any sneaky code, suspicious browser plugins, or AI tools that might be silently leaking access to your data. It covers Claude Code, Codex, Cursor, all of it. Here is why this matters now. For the last six months, hackers have been quietly slipping malicious code into the free building blocks that almost every app in the world is built on. When a developer installs one of these poisoned pieces, the attacker gets a backdoor into everything that developer touches. Including their AI tools and the keys that unlock them. Most security tools defend the finished product. Bumblebee defends the person building it. An independent security researcher read through the entire code and confirmed it is clean. No hidden tracking. No data collection. No backdoors. For two years, AI coding tools shipped with zero security defenses around them. Perplexity just shipped one. Free. Worth installing if you build anything with 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 deeper scans whenever a new supply-chain risk emerges. https://github.com/perplexityai/bumblebee
To get Perplexity Computer and similar tools deeply embedded in enterprises, a continuous investment in security engineering is necessary. What's interesting in the way we're approaching it is putting these tools insde agentic sandboxes and having security workflows run autonomously. Reach out to @kpolley if you're interested in joining and contributing to projects of this nature!
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 deeper scans whenever a new supply-chain risk emerges. https://github.com/perplexityai/bumblebee
MiniMax, a leading open source model and agent, is now powered by Perplexity’s search infrastructure
MiniMax Agent now uses @perplexity_ai Search. We benchmarked 3 AI-native search providers across 700+ agent tasks. Perplexity delivered the best combination of answer quality and snippet density: more useful evidence per token, less irrelevant context, and lower end-to-end
RT Perplexity 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 quality.
RT Kyle Polley Security features that come out of the box with @perplexity_ai Computer: Hyper granular least-privilege access control for your agents This access control includes everything from 3rd party connectors all the way down to the execution run-time environment. Unlike other agent platforms which have access to your entire machine and apps by default, Computer takes the opposite approach and is designed to have the lowest level of access at all times. Fully controlled by you. Best-in-class credential management The agent will never have access to your credentials. We built a network proxy that operates independently from the agent environment and securely validates & hydrates outbound network traffic when necessary. Security Detection & Prevention Computer uses multiple security detection & prevention mechanisms such as BrowseSafe, our open-sourced prompt injection detection system, as well as security sub-agents that scan all Computer-written code for vulnerabilities and fixes them before execution. You do not need to configure or set up any of these security mechanisms! They are just core components of how we designed Computer. Computer should "just work", and that includes security. Much more to come! 🔐
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 API keys.
Perplexity is building one of the most secure scalable agent runtime sandboxes in the market right now. A blog post on how we: 1. Handle proxy API keys for agents securely 2. Run safety detection for all content accessed by agents 3. Encrypt data passed via connectors to agents 4. Decouple storage and compute reliably
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 API keys.
RT NVIDIA Data Center Higher throughput. Lower cost. Same accuracy. @perplexity_ai just put Qwen3 235B into production on NVIDIA GB200 NVL72. Check out how the team made it happen ⬇️
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, not just a training platform.
GB 200s change how one does the prefill and decode disaggregation when serving large MoEs like Qwen. We’ve published details of our stack quantifying the throughput benefits compared to serving on Hoppers.
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, not just a training platform.
RT Perplexity 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, not just a training platform.
alpha
Perplexity Computer + realtime OHLCV data from stock exchanges (included with Computer) + Slack integration
RT Perplexity We’ve developed our own inference engine Runtime-Optimized Serving Engine (ROSE) to serve models ranging from embeddings to trillion-parameter LLMs. With CuTeDSL integrated into our inference engine, Perplexity can build the specialized GPU kernels faster to bring models up to peak performance on NVIDIA Hopper and Blackwell GPUs.
RT Perplexity Finance Search is now available in the Perplexity Agent API. In one tool call, developers can now retrieve licensed financial datasets, real-time market data, and cited web sources for agents that need current, verifiable financial answers.
Perplexity and Computer now allow you to run Deep and Wide Research on sources trusted by doctors and medical professionals like the New England Journal of Medicine, the British Medical Journal, the American Diabetes Association, and so on.
Perplexity and Computer now connect to premium health sources, starting with NEJM and BMJ Group, with 9 more medical journals and clinical databases on the way. Ask health questions and get answers cited from the same sources relied on by hospitals and research institutions.
$AAPL CFO: “Leading AI developers like Perplexity are choosing Mac as their preferred platform to build enterprise-grade AI assistants that power autonomous agents and boost workplace productivity.”
RT Wall St Engine $AAPL CFO: “Leading AI developers like Perplexity are choosing Mac as their preferred platform to build enterprise-grade AI assistants that power autonomous agents and boost workplace productivity.”
RT Nick Davidov It really feels the future of healthcare is here. Perplexity Computer armed with all of my labs, genome, and health data. Constantly on the lookout for new research coming out
We’re rolling out GPT 5.5 as the default orchestrator model for Perplexity Computer. We will be monitoring the user sentiment compared to Opus 4.7 as the default as the rollout expands. Let us know your feedback!
RT Perplexity GPT-5.5 is now available on Perplexity for Max subscribers. GPT-5.5 is also rolling out as the default orchestration model in Computer for both Pro and Max subscribers.
RT Perplexity Kimi K2.6, the new state-of-the-art open-weight model from Moonshot, is now available for Pro and Max subscribers.
We’ve post trained a model on top of Qwen that achieves Pareto optimality on accuracy-cost curves. Unlike our previous post trained models, this model has been trained to be good at search and tool calls simultaneously, allowing us to unify the tool call router and summarization together in one model. The resulting model performs better than GPT and Sonnet in terms of cost efficiency to serve daily Perplexity queries in production. The production model runs on our own inference platform. We’re already serving a significant chunk of our daily traffic with this model and intend to have it serve all of default traffic pretty soon. More research to follow soon on models we’re training and deploying for Comet and Computer.
We've published new research on how we post-train models for accurate search-augmented answers. Our SFT + RL pipeline improves search, citation quality, instruction following, and efficiency. With Qwen models, we match or beat GPT models on factuality at a lower cost.
RT Perplexity We've published new research on how we post-train models for accurate search-augmented answers. Our SFT + RL pipeline improves search, citation quality, instruction following, and efficiency. With Qwen models, we match or beat GPT models on factuality at a lower cost.
As agentic workloads expand, the best orchestration hybridizes centralized servers and CPUs for the ultimate continuity. Especially when you want your AI to use your local apps and messages.
RT Perplexity Claude Opus 4.7 is now the default orchestration model powering Computer. It's also available for Max subscribers on Perplexity web, iOS, and Android.
RT Historic Vids 44 years ago IBM released the first personal computer, establishing the standard for modern personal computing.. we’ve come a long way!
Today we're releasing Personal Computer. Personal Computer integrates with the Perplexity Mac App for secure orchestration across your local files, native apps, and browser. We’re rolling this out to all Perplexity Max subscribers and everyone on the waitlist starting today.
View quoted post“A kid these days could start a multi billion dollar business by using these tools in some way no one has thought about” Enroll and give a shot at securing $1m in funding!
Today we're announcing the Billion Dollar Build. An 8-week competition where teams will use Perplexity Computer to build a company with a path to $1B. Finalists have the opportunity to secure up to $1M in investment from the Perplexity Fund and up to $1M in Computer credits.
RT Lewis Hamilton I’ve been using Perplexity Computer and it’s made me realize that anyone can build their idea. The Billion Dollar build is rewarding exactly this, with up to $1M on the line for whoever can build an entire business with Computer !! Original tweet: https://x.com/LewisHamilton/status/2043427813206671646
Today we're announcing the Billion Dollar Build. An 8-week competition where teams will use Perplexity Computer to build a company with a path to $1B. Finalists have the opportunity to secure up to $1M in investment from the Perplexity Fund and up to $1M in Computer credits.
Perplexity Computer is more reliable than a CPA for filing taxes.
@Yuchenj_UW Perplexity Computer saved me $14k in tax. It found 2 double taxing errors and 2 form filling errors from my $2000-CPA's draft, which CPA fully agreed. In another thread, I let it compute tax from scratch. It's correct to the cents.
RT Perplexity Today we're announcing the Billion Dollar Build. An 8-week competition where teams will use Perplexity Computer to build a company with a path to $1B. Finalists have the opportunity to secure up to $1M in investment from the Perplexity Fund and up to $1M in Computer credits. Original tweet: https://x.com/perplexity_ai/status/2041929222135173466
“This is the closest thing I've come to producing work that looks & feels like it was produced by a well-trained analyst.”
The most exciting outputs I am getting from my experimentation in building an AI-native workflow are not the result of prompts or single workflow skills. The really exciting outputs are a function of orchestrated pipelines, which are a set of sequentially applied skills. The
RT Perplexity Today, we're launching the Secure Intelligence Institute. SII partners with top cryptography, security, and ML teams to advance security research and industry collaboration. It is led by Dr. Ninghui Li at Purdue. https://www.perplexity.ai/secure-intelligence-institute Original tweet: https://x.com/perplexity_ai/status/2039029140758864314
We're deepening our partnership with Samsung by powering the AI on their browser that's pre-installed on 1B+ Samsung devices (100m+ active users), extending the partnership that allows Perplexity to power Bixby and be pre-loaded on all Galaxy S26 devices with alongside Gemini.
Perplexity's search embeddings crossed 1M downloads in less than a month. Remains state-of-the-art with the ability to store more pages per GB with int8 precision.
Today we're releasing two embedding model families, pplx-embed-v1 and pplx-embed-context-v1. These SOTA embedding APIs are designed specifically for real-world, web-scale retrieval. https://pplx.ai/pplx-embed
View quoted postRT Brett Caughran I never found a way to adopt the old version of Perplexity in my investment process, but the things you can one-shot in Perplexity Computer are mind bogglingly good, and it has become a daily driver for me (for the avoidance of doubt, this is not a sponsored post). For example, I always created a guidance credibility analysis, which is a 10-year study of company guidance behavior vs. ultimate results in subsectors: which companies consistently beat/missed/held? These trends can be strong initial indicators of business momentum shifts (i.e. a beat and raise story who just held guidance in Q3 for the first time in 5 years has some information). And you see patterns here that cause you to skeptically question guidance (high or low). Perplexity Computer can one shot this with a 9-page report. Validation is critical, so I may have a different Perplexity Computer space or a Claude CoWork session go and systematically validate, then spot check & validate any key events by hand. But the ability to spin this up quickly, at scale is but one of many emerging capabilities that has me very excited about the very recent agentic capabilities. Original tweet: https://x.com/FundamentEdge/status/2036487922200035577
We're hosting a Perplexity Computer stock pitch competition starting on March 30th for students enrolled in a US undergraduate or graduate program. Students will have 1 week to research, analyze, and pitch a publicly-listed stock, using only Perplexity Computer.
Perplexity Computer Stock Pitch Competition for US undergraduate students. Judged by founders and managing partners of Coatue, Third Point and Silver Lake. March 30th.
We're hosting a Perplexity Computer stock pitch competition starting on March 30th for students enrolled in a US undergraduate or graduate program. Students will have 1 week to research, analyze, and pitch a publicly-listed stock, using only Perplexity Computer.
Nice! Retro Computer
Someone built an old windows 3.1 OS inside the perplexity computer, so we can @AskPerplexity 😭
