OpenAI focuses majority R&D on next-gen GPT models

At the Fellows Forum 2026, Boris Power, head of applied research at OpenAI, disclosed that the company has redirected 80% to 90% of its R&D resources toward GPT-7, GPT-8, and future models. Power stated that these models represent the core of OpenAI's technological value. He explained that while incremental updates within the same generation, such as from GPT-5.1 to GPT-5.2, offer rapid iteration, they are viewed as short-term investments. Major breakthroughs, however, come from the launch of new base models, which shift the team's short-term focus. Power also highlighted that the main challenge in AI assistants is helping users understand and leverage the technology, rather than just improving model quality. Power emphasized that the evolution of product interaction is centered on guiding users to unlock the potential of AI, rather than solely enhancing model performance. This strategic shift underscores OpenAI's commitment to long-term innovation over incremental improvements.