OpenAI‘s GPT-4 held the top position on the Epoch Capabilities Index for approximately one year. OpenAI’s o1 managed a second-longest lead of just over three months. Since Claude 3 Opus removed GPT-4 from the top spot in February 2024, the ranking lead has changed hands seventeen times. The median stay at the top per model is now about seven weeks.
This data highlights how much faster the pace of improvement has become. Rival labs required significantly less time to match GPT-4’s initial performance. No model can maintain a comparable lead anymore because capability jumps occur more frequently. The era that began with reasoning models like o1-preview in late 2024 shows smaller but more rapid transitions between leaders.
* GPT-4’s reign lasted roughly fifty-two weeks
* The current median lead is approximately seven weeks
* There have been seventeen leadership changes since February 2024
