Parallel Cuts Research Time and Cost With GPT-6 Astra

Parallel has integrated OpenAI's GPT-6 Astra model to run developer infrastructure for AI agents performing knowledge work over the web. According to OpenAI's September 22, 2026 announcement, the new model enabled Parallel's agents to research and synthesize labor-market data in half the time and at half the cost compared to prior models.

In one test evaluating six months of data, an agent was tasked with researching six different labor-market statistics across four states over six months. The task required searching multiple websites, collecting information, and compiling the findings into a single report. GPT-6 Astra completed the work in half the time of prior models while cutting code costs by approximately 50% and maintaining research quality.

Parallel noted that the model made more focused searches and took fewer steps to reach useful results. This efficiency allows Parallel to divide research among multiple sub-agents to work simultaneously, reducing the time spent moving through a sequential series of searches.