OpenAI Upgrades GPT-6 Astra Pro in ChatGPT With Faster Responses

The shift toward parallel agents working on separate parts of one task is the more consequential change here, since it moves ChatGPT from a single-response tool toward something closer to a workflow layer. That matters most for teams already routing multi-step work through plug-ins and tools, where latency and coordination, not raw model quality, tend to be the bottleneck. Whether parallel agent output stays reviewable at scale, rather than just faster, is the open question.