OpenAI says most research is already aimed at GPT-7 and beyond

By Splitfeed EditorialLast updated Sep 27, 2026, 10:40 UTC

On September 27, 2026, Boris Power, OpenAI's Head of Applied Research, disclosed that 80 to 90 percent of the company's research is already directed at GPT-7, GPT-8, and models further out. He described improvements made within a single model generation as intentionally short-term bets rather than primary research commitments, meaning the bulk of OpenAI's scientific effort sits several product cycles ahead of anything currently available to the public.

The disclosure matters because it reframes how OpenAI presents the pace of its ambitions. The gap between what OpenAI is publicly releasing and what its researchers are already building toward is, by Power's own account, at least two full model generations wide. Power also shifted attention away from raw capability, arguing that low user awareness of existing AI tools is a more pressing problem than model performance itself. That claim is notable because it turns the usual framing on its head: rather than faster models being the bottleneck, Power suggests adoption and understanding are.

No reactions from the Pro-AI, Anti-AI, or Middle Ground camps have been published yet. Pro-AI voices would typically treat this kind of disclosure as evidence of a healthy and ambitious research pipeline, pointing to the long planning horizons as a sign of organizational maturity. Anti-AI voices would likely raise concerns about the pace of development outstripping safety work, questioning whether research aimed at GPT-8 and beyond is moving faster than the governance structures meant to oversee it. Middle Ground observers would probably focus on Power's user awareness point, arguing it exposes a widening gap between frontier AI development and meaningful public engagement with the technology.

The argument Power's remarks open up will be shaped by what OpenAI shows when GPT-7 eventually arrives, particularly whether its release is accompanied by transparency about the safety and alignment research that ran alongside the capability work. Any concrete timelines or further disclosures from OpenAI leadership about the GPT-7 development schedule would give observers a clearer basis for evaluating how these research priorities translate into products.