OpenAI says most research now targets GPT‑7, beyond
Applied research head Boris Power said 80 to 90 percent of OpenAI's work already goes toward GPT-7 and later models, calling incremental updates like GPT-5.1 to 5.2 short-term bets.
OpenAI's head of applied research, Boris Power, said 80 to 90 percent of the company's research already targets GPT-7, GPT-8 and beyond. That is where "most of the value" comes from, he said at a Fellows Forum event, according to The Decoder.
Power called incremental updates within a single model generation, such as a shift from GPT-5.1 to GPT-5.2, "extremely shortsighted" internally, The Decoder reported. Such releases still rely on specialized training data and help the company iterate quickly in the meantime.
GPT-4 needed careful prompting to get good results, Power said. GPT-5 is easier to use but still needs a lot of user feedback to work well. He said GPT-6 is meant to function more like a capable colleague that only needs a goal set for it.
Power said OpenAI's bigger obstacle is user understanding rather than model quality, according to The Decoder. Most ChatGPT users do not know what the systems can already do, he said, and future versions should better surface those capabilities and anticipate what users need.