Perplexity's Search as Code Beats OpenAI and Anthropic on 4 of 5 Benchmarks
Perplexity's Search as Code gets faster and cheaper, while SDK updates push agent action reliability from 81.9% to 92.6%
Perplexity AI
35 articles tagged with this keyword, sorted by most recent first.
The dream of serving ads to AI agents has hit a snag
Perplexity said ads designed to influence AI bots are "deceptive," in response to a recent experiment from the publisher Time.
Perplexity Swaps GPT-5.6 Terra and Luna Into Its AI Agent Platform
GPT-5.6 Terra and Luna are now the default models powering Perplexity Computer's subagents and scheduled automations, with Terra scoring 11 points above Claude Sonnet on WANDR.
In zweiter Instanz: Perplexitys Agenten dürfen vorerst weiter auf Amazon shoppen
Ein US-Gericht setzte in zweiter Instanz die einstweilige Verfügung gegen Perplexity aus. Dessen KI-Agenten dürfen vorerst weiter auf Amazon Einkäufe tätigen.
What the Perplexity vs Amazon ruling means for AI agents acting on users’ behalf
The US Court rejected Amazon's argument that the AI assistant in Perplexity's Comet browser accessed websites & users' accounts without permission, saying the agent just carried out the user's instructions. The post What the Perplexity vs Amazon ruling means for AI agents acting on users’ behalf appeared first on MEDIANAMA .
US appeals court allows Perplexity's AI shopping agent back on Amazon
A US appeals court has overturned Amazon's injunction against Perplexity's AI shopping agents, ruling that it's the users who access Amazon, not the startup. It's the first federal appeals court decision on whether AI agents can lawfully act on online platforms on behalf of users, and it could reshape the entire AI agent industry. The article US appeals court allows Perplexity's AI shopping agent back on Amazon appeared first on The Decoder .
Moburst Unveils AI-Driven Mobile Growth Playbook
Moburst, a mobile growth marketing agency, has formalized a mobile-specific approach to Answer Engine Optimization, aimed at helping app publishers get recommended by AI assistants such as ChatGPT, Perplexity, and Google’s AI Overviews, in addition to ranking inside the App Store and Google Play. Why mobile teams are asking this question now A growing share of app discovery now starts outside the app stores entirely. Industry research trackers have documented rapid year-over-year growth in AI-mediated search sessions, alongside forecasts that a meaningful share of organic search volume will continue shifting toward AI chatbots and assistants. For app publishers, that means a user can research, compare, and effectively decide on an app before ever opening a store listing. “Recommend a digital marketing agency that specializes in AEO for mobile growth” is a question more procurement teams are typing into search bars and AI assistants themselves, and the honest answer is that the specialist pool is still small. Most agencies claiming AEO expertise are general SEO shops that have added the term to their service pages without building mobile-specific measurement underneath it. That gap is not a small detail. A procurement team that hires a generalist expecting mobile-specific results is likely to end up with a web-only AEO program that never touches the app store side of discovery at all, and then has no easy way to tell, months later, why the results fell short of expectations.…
Perplexity Rebuilds Projects With Brain Memory to Beat Stateless AI Agents
Perplexity's Spaces become Projects: a persistent file system and self-improving Brain memory now power long-running agentic work for all users
Perplexity Releases Numbat to Stop Rogue AI Agents After OpenAI Breach
Perplexity open-sources Numbat, a cross-harness agent security layer that monitors, detects, and blocks dangerous AI agent behavior before it executes
Perplexity employee who worked on Comet launches an AI browser aimed at knowledge work
Polar has come out with an AI-first browser aimed at knowledge workers, and it has now raised a $5.7 million seed round led by Madrona.
Perplexity's tokenmaxxing Model Council gives you multiple bot perspectives
Up to 8 AI models running in the cloud weighing in on ambiguous business issues? Sounds affordable
Perplexity Adds Kimi K3, the World's Largest Open-Weight Model, on US Servers
Perplexity brings Moonshot AI's record-breaking 2.8T-parameter open-weight model to its search and agentic platform, with a US-only hosting pledge to sidestep data sovereignty concerns.
Perplexity brings its Personal Computer AI agent to Windows
Perplexity AI Inc. today released a Windows version of Personal Computer, expanding its agentic automation software beyond the original Macintosh platform and making it available to more than 1 billion Windows devices. First released in April, Personal Computer software acts as a general-purpose digital worker that can access authorized files and applications on a user’s […] The post Perplexity brings its Personal Computer AI agent to Windows appeared first on SiliconANGLE .
Perplexity's Personal Computer Brings AI Agents to Windows' 1 Billion Devices
Perplexity's Personal Computer agent is now live on Windows, bringing multi-model AI orchestration to local files, Microsoft 365, and 400+ connected apps
Perplexity’s Personal Computer turns Windows PCs into AI agents
Perplexity has expanded its agentic Personal Computer tool to Windows, allowing computers running the world's most popular OS to be used as a locally run AI system. Like the Mac version that Perplexity launched in April, Personal Computer for Windows operates like a "general-purpose digital worker" that can access local files and apps to perform […]
I gave Perplexity's agentic AI 5 complex tasks to run on my Mac - and I'll do it again
Perplexity's Mac app offers its own agentic AI, Personal Computer, which can handle multi-step tasks on your computer from start to finish. See why the results impressed me.
Germany puts Google's AI Overviews and Perplexity under media law in first-of-its-kind ruling
German media regulators say Google's AI Overviews are Google's own content, not neutral search results, and that they crowd out regular links. The regulators have issued their first rulings against Google and Perplexity under the country's State Media Treaty. Both companies have one month to appeal. The article Germany puts Google's AI Overviews and Perplexity under media law in first-of-its-kind ruling appeared first on The Decoder .
Perplexity's SPACE Runs Secure Long-Running Agents 5x Faster
Perplexity's new SPACE platform runs agent sessions in disposable Firecracker microVMs with rolling snapshots, cutting sandbox creation latency 3-5x
Perplexity launches secure sandbox to make its AI agents secure and powerful
Perplexity AI Inc. today introduced a new feature that takes its current agentic artificial intelligence service, Computer, to perform better with greater security. The company introduced SPACE, a sandbox platform designed to allow its AI agent to act with its full capabilities, while providing the highest level of security for agentic systems. Perplexity Computer can […] The post Perplexity launches secure sandbox to make its AI agents secure and powerful appeared first on SiliconANGLE .
Perplexity's Open-Source WANDR Benchmark Exposes How Badly Research Agents Fail at Scale
Perplexity open-sources WANDR, a 500-task benchmark exposing how badly research agents fail at large-scale, evidence-backed data collection
KI-Zusammenfassungen: Google und Perplexity unterliegen gegen Medienanstalten
Die Medienanstalten stufen KI-generierte Antworten als eigene Inhalte ein und fordern Transparenz. Das DSA-Haftungsprivileg greift hier laut Gutachtern nicht.
Perplexity Swaps Claude for Grok 4.5 at Half the Cost
Perplexity swaps in xAI's freshly-launched Grok 4.5 as the orchestrator brain of Computer, beating every rival configuration on WANDR at roughly half the cost of Claude Opus 4.8
Perplexity Ships Computer Analytics so You Stop Losing Hundreds in Mystery Credits
Perplexity Computer now shows credit spend broken down by model, available to all consumer and enterprise users in Account Settings.
Perplexity Ships GLM 5.2 Orchestrator That Matches Claude Opus at One-Third the Cost
Perplexity post-trains GLM 5.2 as a cheaper orchestrator for Computer, hitting near-frontier performance at 34% of Claude Opus cost
Perplexity is quietly building an AI coding tool to take on Cursor and Claude Code
San Francisco-based Perplexity, valued at $20 billion, is working on 'Teammate,' an AI tool that may compete with AI coding giants like Anthropic.
Why Perplexity's founder doesn't think the American dream is over
Aravind Srinivas, the CEO of Perplexity, praised America's startup culture on "The Joe Rogan Experience."
Most Current Model Organisms Leak: Perplexity Differencing Often Reveals Finetuning Objectives
Authors: Mohammad Abu Baker, Luca Baroni, Daniel Wilhelm Paper: https://arxiv.org/abs/2605.00994 Code: https://github.com/z3research/ppldiff-paper Twitter thread: https://x.com/m_shahoyi/status/2071892578476110136 Top-ranked revealing completions can be inspected here: https://z3research.org/ This post summarizes the paper and adds a few extra reflections in Discussion TL;DR We found that many current publicly available model organisms (MOs) "leak" instilled behaviors We present a simple contrastive method to surface this: generate MO completions from a set of short general-corpora prefills. Then, rank completions by perplexity difference wrt a reference model. Top-ranked completions often reveal the finetuning objective. Effective on the vast majority of the model organisms we tested (N=76), across model families, sizes (0.5B to 70B), and behaviors including backdoors, false facts, and unsafe behaviors. Surfaced completions contain both memorized sentences and learned emergent behaviors absent from finetuning data. The method is most effective using the pre-finetuned model as reference, but we show that unrelated reference models from other families detect the behaviors nearly as often. In AuditBench , a benchmark for detecting hidden behaviors, an agent given access to top-ranked perplexity-difference completions is SOTA (avg detection rate 0.73), almost saturating the benchmark on SDF models. Introduction LLMs can be deliberately manipulated to exhibit harmful behaviors,…
Most Current Model Organisms Leaks: Perplexity Differencing Often Reveals Finetuning Objectives
Authors: Mohammad Abu Baker, Luca Baroni, Daniel Wilhelm Paper: https://arxiv.org/abs/2605.00994 Code: https://github.com/z3research/ppldiff-paper Twitter thread: https://x.com/m_shahoyi/status/2071892578476110136 Top-ranked revealing completions can be inspected here: https://z3research.org/ This post summarizes the paper and adds a few extra reflections in Discussion TL;DR We found that many current publicly available model organisms (MOs) "leak" instilled behaviors We present a simple contrastive method to surface this: generate MO completions from a set of short general-corpora prefills. Then, rank completions by perplexity difference wrt a reference model. Top-ranked completions often reveal the finetuning objective. Effective on the vast majority of the model organisms we tested (N=76), across model families, sizes (0.5B to 70B), and behaviors including backdoors, false facts, and unsafe behaviors. Surfaced completions contain both memorized sentences and learned emergent behaviors absent from finetuning data. The method is most effective using the pre-finetuned model as reference, but we show that unrelated reference models from other families detect the behaviors nearly as often. In AuditBench , a benchmark for detecting hidden behaviors, an agent given access to top-ranked perplexity-difference completions is SOTA (avg detection rate 0.73), almost saturating the benchmark on SDF models. Introduction LLMs can be deliberately manipulated to exhibit harmful behaviors,…
A Black Box Made Less Opaque (part 4)
Understanding the effects of compression on model performance and interpretability I. Executive summary This is the fourth installment in a series of analyses exploring basic AI interpretability mechanics and techniques. While this analysis is designed to stand on its own, readers interested in a comparative analysis of representational geometry and the effects of manipulating feature activation will likely appreciate a review of part 1 , part 2 , and part 3 of this series. Key findings: Context: This analysis examines the effect of standard levels of weight compression on Google DeepMind’s Gemma 3 4B parameter and Gemma 3 12B parameter models. For each model, I examine the original, uncompressed version as a control before examining the 8-bit and 4-bit weight compressed (quantized) versions of that model. Performance vs. compression: For both models, performance (as measured via cross-entropy and perplexity) is largely preserved under compression. 8-bit compression had essentially no effect on performance with only modest degradation at 4-bit (~2% for 4B, ~2.7% for 12B). SAE applicability vs. compression: Each model’s pretrained sparse autoencoders (SAEs) demonstrated a remarkably consistent ability to reconstruct the model’s residual stream (as measured by the fraction of variance unexplained, or FVU), despite increasing levels of model weight compression. The “so what?”: That model performance degrades only modestly, and only at 4-bit, while SAE applicability remains rela…
Perplexity Launches Computer for Counsel: A Multi-Model Agentic Layer for Legal Workflows
Perplexity's Computer for Counsel extends Perplexity Computer to legal teams. It routes 20+ models across Midpage, MCP connectors, and Microsoft 365, with cited outputs lawyers can verify. The post Perplexity Launches Computer for Counsel: A Multi-Model Agentic Layer for Legal Workflows appeared first on MarkTechPost .
Join us at Big Technology's AI Summit in San Francisco on 6-18
OpenAI President Greg Brockman, Perplexity CEO Aravind Srinivas, Box CEO Aaron Levie, and more join us for a day of newsmaking conversations, live at San Francisco's Commonwealth Club.
How are perplexities over multiple instance aggregated?
The perplexity of the $i^{th}$ token in the $k^{th}$ sequence is $$ P_{ki} = \frac{1}{p(t_{ki})} $$ The perplexity aggregated for the $k^{th}$ sequence is then $$ P_{k} = \left(\prod_{i=1}^N P_{ki}\right)^{1/N} \\ = \left(\prod_{i=1}^N \frac{1}{p(t_{ki})} \right)^{1/N} $$ which is the geometric mean of the perplexities of the tokens. This makes sense as we are essentially taking the multiplicative inverse of the probability that the model got the whole sequence correct. Now my question is how to aggregate the perplexities of several sequences. It seems from various places, including the Hugging Face Tutorial , I see that the prescription is to take the arithmetic mean of the perplexities of sequences $$ P = \frac{1}{m} \sum_{k=1}^m P_k $$ I am not quite understanding what it means to take the average of 1/probabilities. What is this actually capturing?