Parallel, a developer infrastructure company for AI agents that perform knowledge work over the web, has significantly reduced research time and costs using GPT-6 Astra. The company's tools support applications ranging from web grounding for voice agents to research for financial and legal institutions. Previously, high-quality answers for long-running tasks required larger models with extended reasoning, consuming more time and resources. With GPT-6 Astra, Parallel achieved the same research quality in half the time with roughly 50% code cost reduction.

In a test involving six labor-market statistics across four states over six months, GPT-6 Astra completed the work twice as fast as prior models. The agent searched multiple websites, collected information, and compiled findings into a single report. According to Devin Gupta, Member of Technical Staff at Parallel Web Systems, Astra issued more targeted search queries and focused better on the ultimate task by incorporating its world knowledge.

The increased efficiency enables Parallel to divide research among multiple agents. GPT-6 Astra can delegate specific tasks to sub-agents, allowing simultaneous work and reducing time spent on sequential searches. This provides a better path from complex questions to researched answers with less waiting, lower costs, and more capacity for demanding research tasks at scale.