Pervaziv AI Updates Cortex Discover and Cortex Cloud with Faster Intelligence and Broader Campaigns
Discover 1.5 improves startup, responsiveness and resource efficiency while Cloud 1.5 expands managed validation into
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.
Discover 1.5 improves startup, responsiveness and resource efficiency while Cloud 1.5 expands managed validation into coordinated campaigns across software repositories.
San Francisco, CA (PRUnderground) October 27th, 2026
Pervaziv AI today announced 1.5 updates for two distinct products in the Cortex platform, Cortex Discover and Cortex Cloud, strengthening both the interactive and execution sides of AI assisted engineering.
Cortex Discover 1.5 continues the company’s performance program for agentic browsing with faster startup, near memory parity with Chrome in controlled workloads, responsive interaction, full speed rendering and improved media efficiency. Cortex Cloud 1.5 addresses a different stage of the engineering lifecycle by expanding managed execution from individual repository validation into coordinated campaigns across multiple repositories, while strengthening execution history, reproducibility, diagnostics, workload isolation, performance visibility and estimated AI cost.
Together, the updates reflect a broader direction for Cortex: AI should remain fast and practical where people interact with it, and when that work moves into execution, organizations should be able to validate it across the software systems that actually matter.
Fast where work begins. Evidence when work moves forward.
The Benchmark Had to Survive the Next Test
The latest Cortex Discover work follows a performance program that has deliberately become more demanding over time. Rather than optimizing around a favorable benchmark and moving on, Pervaziv AI has continued testing whether the performance characteristics of Cortex Discover hold as the product evolves.
The approach started with a simple principle: we began by challenging our own strongest result.
Earlier Cortex Discover research led Pervaziv AI to revisit and refine previous performance conclusions as testing became more controlled. When tighter measurement changed the result, the company changed the headline rather than preserving the stronger claim. That philosophy continues with Cortex Discover 1.5.
The objective is not to claim that one browser wins every possible workload. It is to determine whether an agentic browser can add intelligence while preserving the responsiveness and efficiency expected from everyday browsing. The latest results provide another step in that direction.
Faster Startup Without Heavier Everyday Browsing
Startup remains one of the clearest measured advantages for Cortex Discover. In the latest controlled comparison, Cortex Discover 1.5 delivered 14.8% faster native startup than Chrome, while warm startup was also 6.1% faster.
Those results matter because an agentic browser may become part of technical research, documentation, enterprise applications, development workflows and longer running AI assisted activity. Intelligence should not introduce unnecessary friction before the work even begins.
At the same time, Cortex Discover remained effectively at memory parity with Chrome across single tab, multi tab and active workloads. That is an important result for a product designed to keep AI available throughout everyday work.
Persistent AI introduces a different resource challenge from an occasional assistant. A browser already manages pages, rendering, media, scripts, applications and extensions. An agentic browser adds context awareness, conversations, AI workflows and connected intelligence. The goal is not to pretend that intelligence has no resource cost, but to keep that cost disciplined.
For Cortex Discover, performance is part of the AI experience.
Responsive Browsing Still Comes First
Startup and memory are only part of browser performance. A browser also has to remain responsive after it opens, especially when users are moving between research, web applications, documentation, visual content and AI assisted workflows.
In the latest controlled testing, Cortex Discover maintained responsive interaction performance while continuing to provide its integrated AI experience. Rendering remained at 60 FPS in the measured workload, matching the browser baseline, while repeated interaction testing showed competitive performance across the evaluated scenarios.
These results reinforce an important design goal for Discover: the AI experience should coexist with ordinary browser behavior rather than forcing users to choose between intelligence and responsiveness. Cortex Discover is built around the idea that the browser itself can become an intelligent work environment while still being expected to perform well at the fundamentals of browsing.
Media Efficiency Shows Another Direction
Modern browser workloads extend well beyond static pages. Video, dashboards, rich applications, visual interfaces and browser based development tools are now routine parts of professional work, making media efficiency another important part of the overall experience.
Cortex Discover 1.5 showed encouraging results in controlled media testing. A compact deterministic H.264 workload measured 12.8% lower CPU usage for Cortex Discover while maintaining near equivalent output.
The measurement represents a specific controlled workload rather than a universal claim across every video service or device, but its significance is the direction it supports. An intelligent browser still has to be an efficient browser, and as AI becomes more deeply integrated into everyday computing, resource efficiency becomes part of product quality.
Cortex Cloud 1.5 Tackles a Different Problem
Cortex Cloud 1.5 begins where the interactive experience ends. Generating or reviewing a change is not the same as proving that the resulting software works across a real product.
Production applications frequently span multiple repositories. A single feature may touch an application, backend service, API, shared library, infrastructure configuration and deployment workflow. Testing those components only as isolated projects can leave important questions unanswered, including whether the broader change actually succeeded, whether one repository improved while another failed, and how much execution capacity or AI spending the work consumed.
Cortex Cloud 1.5 introduces multi repository benchmark campaigns designed to make those questions easier to answer.
From Individual Jobs to Connected Campaigns
Teams can organize related repository evaluations into a shared campaign, follow active work, stop campaigns when priorities change and return to previous evaluations without manually rebuilding the original context. A campaign can represent a release candidate, model comparison, development agent evaluation, dependency update, infrastructure change or broader modernization project.
Individual repositories retain their own validation results, while the campaign adds the product level context. This allows teams to understand not just whether one job passed, but how the larger set of software performed.
That becomes increasingly valuable as AI generated changes cross service boundaries. A modification that appears successful in one repository may introduce an incompatibility elsewhere. A new model may improve completion on one project while consuming more time or resources across another group of workloads. Campaign based evaluation makes those patterns easier to see.
History Becomes Engineering Evidence
Cortex Cloud 1.5 also strengthens the relationship between validation results and the software that produced them. A stored pass or fail status has limited value if teams cannot later determine exactly what was evaluated.
Cloud 1.5 preserves stronger connections among source identity, repository versions, validation inputs, execution records and resulting evidence. Teams can revisit an earlier campaign and understand which source state was tested rather than relying only on a branch name or moving label.
That turns historical campaigns into more than troubleshooting records. They become comparable engineering evidence that can help organizations evaluate models, compare releases, identify recurring execution patterns and understand whether newer workflows are actually improving outcomes.
For example, organizations evaluating a new coding model can look beyond whether it generated working code and instead compare completion rates, validation time, resource consumption and estimated AI cost across representative projects.
Reliability Without Hiding Failure
Managed automation becomes useful only when its results remain understandable. Real repositories vary significantly in how they prepare and run, and dependencies, assets or required tooling may be unavailable before validation even begins.
Cortex Cloud 1.5 improves preparation so supported projects can reach meaningful execution more consistently. The system can identify required resources before validation begins and apply bounded recovery when a recognized preparation condition can be corrected safely.
Preparation, however, remains distinct from product validation. A successfully prepared environment does not become a successful repository test, and a preparation repair does not silently change the selected validation suite. When a blocking condition cannot be addressed safely, Cortex Cloud stops and reports it.
This balance is important for trustworthy AI assisted engineering. Automation should remove unnecessary operational friction without converting unresolved problems into reassuring results.
Better Diagnostics, Stronger Isolation
Cloud 1.5 also improves how unsuccessful work is explained. Distributed validation can fail during environment preparation, dependency retrieval, builds, workload startup or the selected tests themselves, and treating every failure as an undifferentiated error makes engineering teams spend more time reconstructing what happened.
Cortex Cloud retains more useful context around execution stages and operations, helping teams separate repository problems, external conditions, policy restrictions and environment issues.
At the same time, the release strengthens workload separation as more jobs run concurrently. Build data, cached dependencies, execution resources and network activity are more carefully separated between jobs. Resources are associated more closely with the work that owns them, while recovery processes can better identify state created by interrupted execution.
Isolation matters for reliability and security, but it also matters for measurement. If one workload unintentionally influences another, apparent improvements may reflect leftover state rather than the software under evaluation. Stronger boundaries create more confidence that campaign results describe the work that actually ran.
Performance and AI Cost in the Same View
AI assisted development introduces another operational question: what did the result cost to produce?
A faster model is not automatically better if its output creates more failures downstream, and a more capable model may not be the right choice for every project if it substantially increases cost without improving validated outcomes.
Cortex Cloud 1.5 brings validation results, turnaround time, resource consumption and estimated AI spending into a more connected operating view. Developers can investigate technical failures while engineering leaders evaluate broader efficiency, making campaigns useful for model comparisons, platform planning, workflow evaluation and decisions about where AI delivers the greatest engineering value.
The goal is not to reduce engineering quality to one score. It is to expose enough context for organizations to make informed tradeoffs.
Two Products, One Broader Direction
Cortex Discover and Cortex Cloud remain distinct products solving different problems. Discover brings Cortex directly into an intelligent browsing experience designed around fast interaction, context and everyday work. Cloud provides managed execution for work that needs broader validation, reproducibility, workload separation and reviewable evidence.
The connection is the engineering journey. Work can begin with investigation and reasoning in the browser, continue through AI assisted development, and then move into managed execution where broader checks can be performed and results retained for review.
“AI should stay fast where people work and produce evidence when work moves into execution. Discover 1.5 and Cloud 1.5 push both sides forward,” said Anoop Jaishankar, Founder and CEO of Pervaziv AI.
Irina Calciu, Founding Systems Architect at Pervaziv AI, added, “Cloud 1.5 turns validation from isolated jobs into repeatable campaigns across real products, with clearer evidence, stronger isolation and better operational context.“
The two releases build on recent Cortex work across agentic browsing, managed execution, security, workflow continuity, decision intelligence and specialized AI. Their message, however, is intentionally practical: more AI capability should not automatically mean more friction, and more AI generated work should not mean less evidence.
Faster Intelligence. Broader Execution.
Cortex Discover 1.5 strengthens the interactive side of Cortex with 14.8% faster native startup, responsive browser performance, near memory parity and improved media efficiency. Cortex Cloud 1.5 strengthens the execution side with multi repository campaigns, stronger history, clearer diagnostics, improved isolation, reproducibility and better visibility into performance and estimated AI cost.
The products solve different engineering problems, but together they support an increasingly connected path from intent to execution to evidence.
Cortex Discover 1.5 keeps intelligence moving. Cortex Cloud 1.5 helps prove what happens next.
