Parallel Web Systems

Two new models join our Search Capability Leaderboard today: - Pareto 26.9 from Unbiased AI is the first blended model we've ranked on this leaderboard, combining multiple models behind a single, task-routed endpoint - GPT-6.1 Sol from OpenAI is an upgrade to the recently launched GPT-6 Sol Both models are on the Pareto frontier of quality vs. cost. Notably, Pareto 26.9 is extremely cost-efficient, delivering near Fable 5.1 quality at 9x lower cost. GPT 6.1 Sol significantly improves on 6, with quality and cost gains that bring it in line with the latest generation of frontier models at the mid-tier.

Are you overpaying for web search in your agents? Find out with our new calculator: https://lnkd.in/gHnQh7GZ

Parallel is now live in Lovable. Creating intelligent apps powered by the web has never been easier. Just tell Lovable to use Parallel. Here are some ideas we love: - An app for a contractor to monitor for new home sales in the area - A trend researcher for a local boutique, scanning the web for weekly social trends to jump on - A local events tracker to help manage restaurant staffing Get started in Lovable: https://lnkd.in/dBsyYwta

Not a lot of people can see the future. Even fewer can build it. This is the best explanation (for both customers and potential colleagues) of "why Parallel". Great discussion Parag Agrawal and Harry Stebbings. Excited to be building the next wave of the AI evolution. https://lnkd.in/gve3_4G7

Parallel Web Systems builds category-leading web research infrastructure for AI agents. Their APIs let agents search, read, and reason across the web, and return cited, structured answers at scale. But even great web research can't always get agents everything they need. Parallel gets that. That's why, alongside their web research APIs, they built Data Connectors to bring specialized data into the same research. Faraday is proud to be a Data Connector launch partner. If your team already uses Faraday, you can connect it to Parallel's Task and Responses APIs, and each run can draw on Faraday's consumer context alongside the open web. And if you're using Parallel, check out what Faraday's 1,400+ attributes on about 240 million U.S. adults can add to your agents. Read Parallel's announcement: https://lnkd.in/ekJYkHPQ

Parallel started with the belief that agents would use the web 1,000x more than humans and that they would need new infrastructure and business models. The shift is happening faster than we expected. Watch Parag Agrawal discuss with Harry Stebbings on the latest episode of 20VC: YouTube: https://lnkd.in/gXKPDR3E

GPT-6 Sol and Luna are now live on the Search Capability Leaderboard, our composite benchmark for measuring how well models perform web research tasks with Parallel. 🌒 GPT-6 Luna is Pareto-optimal, improving both on cost and quality compared to the previous generation and leading our efficiency leaderboard ☀️ GPT-6 Sol delivers similar capabilities to GPT-5.6 at a ~38% lower cost Explore the leaderboard: https://lnkd.in/g-d4ttMJ

Managed Deep Agents from LangChain launched today with built-in web search powered by Parallel. Add one MCP server to your tools file. LangSmith manages the credentials, runs the calls, and traces every search: no extra account or API key. Bonus: It’s free during the public beta of Managed Deep Agents! Learn more: https://lnkd.in/gc79gXrQ