AIs will soon use the internet far more than humans ever have. At Parallel, we are building for the web’s second user.
Our API is the first to surpass humans and all leading AI models (including GPT-5) on deep web research tasks.
Parallel’s performance on independent benchmarks:
• BrowseComp: 58% accuracy vs GPT-5 (41%) vs humans (25%)
• DeepResearch Bench: 82% win rate over reference vs GPT-5 (66%). Parallel beats GPT-5 head-to-head 74% of the time.
For $0.10, we exceed what takes humans 2 hours on BrowseComp.
The open web’s greatest chapter is just beginning.
The web’s second user
AIs will use the web far more than humans ever have. As much as we might anthropomorphise AIs, they operate differently from humans on the web. They can retrieve and consume thousands of documents in a call, or want just one discrete fact. The infrastructure designed for human use can’t serve the needs of AIs.
At Parallel, we are building for the web’s second user. We are creating systems and infrastructure for AIs to use the web effectively for completing complex tasks. Our suite of products include low-level search tools and deep research APIs that can complete hours of human work on the web in minutes.
State-of-the-art deep web research, available to all as an API
Today, we’re announcing the only AI system to outperform both humans and leading AI models like GPT-5 on the most rigorous benchmarks for deep web research. Our APIs are now broadly available, bringing production-grade web intelligence to any AI agent, application, or workflow.
Parallel already powers millions of research tasks daily, across ambitious startups and public enterprises. Some of the fastest growing AI companies use Parallel to bring web intelligence directly into their platform and agents. Public enterprises automate traditionally-human workflows exceeding human-level accuracy with Parallel. Coding agents rely on our search to find docs and debug issues.
We outperform humans and all leading AI models on deep research tasks
Two of the hardest independent benchmarks in AI web research, BrowseComp (built by OpenAI) and DeepResearch Bench, show Parallel outperforms humans and all leading AI models.
About Parallel
The open web is a miracle. Anyone can publish, learn, and collaborate. It’s the closest thing to humanity’s living memory. This open ecosystem fueled today’s AI breakthroughs.
And it’s about to face its biggest challenge yet: the primary user is shifting from humans to AIs. Soon, AIs will use the web far more than humans ever have.
Human web use is narrow: clicks, scrolls, short searches as inputs; a few pages at a time, all served in seconds. AIs vary widely: instant responses or hours of processing, single facts or entire databases.
Business models built on human attention can’t serve AIs. By default, the web trends toward zero-sum: paywalls, gated APIs, and private data silos everywhere, undermining the very miracle that sparked the AI revolution.
We need to build a new Programmatic Web specifically for AIs: declarative, composable layers built around reasoning and computation, verifiable provenance, and open markets.
Unified data, compute, reasoning. Engineered to operate as one so that outputs are actions and insights, not just documents.
Declarative interfaces. Designed so AIs state what they need, not how to get it. The infrastructure determines how. Pull turns to push.
Transparent attribution. Built so every source and insight is tracked and credited. Contributions become measurable and transparent.
Open, value-based markets. Incentivized so participants earn based on value they add. Staying open wins, not due to virtue, but because it’s economically superior.
This isn’t about preserving the old web. It’s about unlocking what comes next: AIs solving complex problems, accelerating discovery, creating things we can’t yet imagine.
The choice is binary: we build the open web for its second user, or it fractures beyond repair. At Parallel, we are building for abundance.
Travers Nisbet, Cofounder at Parallel
We started Parallel Web Systems because the primary user of the web is changing: from humans to AIs.
When the user changes, everything changes: the interfaces we build, the infrastructure we need, and the markets that grow around it.
At Parallel, we are building for AIs, the web’s next user. Web search is still bound to infrastructure built for the human-first web. This is the primary bottleneck that we aim to unlock.
Today, we’re excited to announce our state-of-the-art search system is the first to surpass humans and all leading AI models (including GPT-5) on deep web research tasks. Our API is available now for developers and companies building the most advanced AI systems, apps, and agents.
Parag Agrawal, Cofounder at Parallel
We already power millions of research tasks every day, across ambitious startups and public enterprises. Some of the fastest growing AI companies use Parallel to bring web intelligence directly into their platform and agents. A public company automates traditionally-human workflows exceeding human-level accuracy with Parallel. Coding agents rely on our search to find docs and debug issues.
And today, we launched our Deep Research API – it’s the first to outperform both humans and all leading models including GPT-5 on two of the hardest benchmarks.
Todd Jackson, Partner at First Round Capital
10 years ago, Parag Agrawal and I were working together to introduce ranking to Twitter’s timeline for hundreds of millions of consumers. Now he’s started Parallel Web Systems to build infrastructure for a very different kind of user: AI agents. And in a full circle moment, we’re able to work together again, as First Round Capital invested in his seed round last year alongside Khosla Ventures and Index Ventures.
Parag and I had a chance to sit down for a great conversation last week (his first interview since leaving Twitter!). It was a fascinating reflection on what he learned and what he’s doing differently.
We dig into:
– The transition from CTO → CEO → Founder
– What he was *really* thinking when all the Twitter drama was going down
– Why he chose to be a founder (vs. take another big role)
– How he landed on the idea for Parallel
– What fundamentally changes when you’re designing for AIs vs. humans
– How he thought about fundraising and building his early team
– How working with customers like Clay shaped Parallel’s earliest products
– The case for building “slow” APIs, and the advice he got from Patrick Collison
– His takes on the modern agent stack and how evals need to evolve
And much more!
Immensely proud to be able to partner with Parag and the entire Parallel team. They’ve made incredible progress in a short amount of time and I can’t wait for more folks to start playing around with the APIs they’ve built.
