October was among our biggest months yet at Parallel. Here’s a recap of everything we shipped: | | Parallel Decisions API | Our new API combines the speed and convenience of models like Jev with the collective knowledge of the live web. Developers can now ask questions about anything: not just what is in their context, but the web as a whole. Web grounding significantly expands the usefulness of decision models in agent workflows that need fast answers. | We also published PWS-Decisions, an open benchmark of 300 questions. The Decisions API scored 76.2%, ahead of Perplexity at 50.9% and OpenAI at 41.7%. It’s free for the next 7 days on OpenRouter. | To learn more, read our latest blog on the Decisions API. | | SOTA web search for Finance | Parallel for Finance upgrades our index and ranking for financial agents across all of Parallel's APIs, with no new plan or subscription. On FinSearchComp's hardest tier, Parallel is the most accurate search API at up to 2x lower cost than Exa and Perplexity. Rogo, Kepler, and Bretton already run Parallel in production, and our new Rogo case study shows how Rogo runs hundreds of thousands of requests a day on Parallel. | To learn more, read our blog on Parallel for Finance. | | See the whole picture: A brand campaign built with Parallel | For our first out-of-home campaign, we used Parallel's own research APIs to build six collages, each grounded in a real Task API run with part of the response inside the artwork. See the campaign for yourself across New York City, San Francisco, and London. | Read about the creation of the campaign here. | | EU Data Residency | Businesses can now use Parallel to serve European customers whose vendor reviews or contracts require EU processing. When you send a request to our EU endpoint, we process and serve it in the European Union, and we don’t retain the request or response content. To use it, send Search and Extract requests to eu.parallel.ai instead of api.parallel.ai. | To learn more, reference our EU data residency docs. | | More news | | | |
|