OpenAI's Reasoning Models Rapidly Improving, Scaling Limits Unclear

OpenAI's reasoning models like o3 are rapidly improving, but it's uncertain how much further they can scale, with current training compute likely below the frontier but not drastically behind, suggesting potential for continued rapid scaling in the near term before growth slows.

• Reasoning models like OpenAI's o3 have seen rapid improvements but it's unclear how much further they can scale
• Current reasoning training compute is likely below the overall frontier but not orders of magnitude behind, suggesting potential for continued rapid scaling in the near term
• If reasoning training scales up to the overall compute frontier within a year or so, its growth rate will slow to around 4x per year in line with overall compute growth